Search

The Multi-Swarm approach allows the use of multiple configurations between two or more populations of particles, where each one can present different approaches (e.g. lbest, gbest, Unified, Guaranteed-Convergence) directed towards improving the optimization process. This article presents a proposal for local/global stochastic interconnection applied to the context of the Multi-Swarm algorithm, as well as for incrementing a local search method for refining previously obtained solutions. Two proposals are introduced for this new Multi-Swarm PSO (MSO). The first one is the inclusion of "counterpart particles", which establishes a sub-topology between inter-swarm particles, accessed by migration and evaluability rules. The other involves using customized crossover operators and is based on the BLX scheme (Blend Crossover) with direction information used as a reference for establish a subspace search around the particles. Performance and robustness of the new approaches were assessed by ten constrained engineering design optimization problems (COPs), as is compared to other solutions already published in the scientific literature. Results indicate significant performance improvements for all 10 COPs when compared to concurrent-based MSOs. By making available new references from other swarms, the counterpart particles approach tends to improve the optimization process in the search space, while an intermediate layer of local search based on a modified directed BLX crossover should provide an extra search around the particle, and thus, refining previously obtained solutions.

Because data centers are energy-intensive industries, data centers have made significant efforts to improve energy efficiency to save energy cost and utilize renewable energy to contribute to environmental sustainability. In addition, data centers have been considered as a potential application of demand response because they can possibly adjust energy consumption to process workload in response to time-of-use prices and intermittent renewable generation. Quality of service (QoS) should be guaranteed at the desired level to respond to users' requests with their satisfaction in an effort to achieve energy-efficient and sustainable data center operations. Given this context, this study intends to develop a two-stage stochastic program that can be used to investigate optimal data center operations based on demand response that results in minimum energy cost with QoS guarantee against the stochastic workloads. Then, an additional chance constraint and risk constraint are considered as means of managing the level of QoS guarantee, and comprehensive numerical experiments are conducted with various parameter settings to evaluates the impact of the QoS guarantee with different levels on overall performance. The results show that the risk associated with the QoS guarantee for data centers needs to be properly managed to improve energy efficiency.

A colored traveling salesman problem (CTSP) as a generalization of the well-known multiple traveling salesman problem utilizes colors to distinguish the accessibility of individual cities to salesmen. This work formulates a precedence-constrained CTSP (PCTSP) over hypergraphs with asymmetric city distances. It is capable of modeling the problems with operations or activities constrained to precedence relationships in many applications. Two types of precedence constraints are taken into account, i.e., 1) among individual cities and 2) among city clusters. An augmented variable neighborhood search (VNS) called POPMUSIC-based VNS (PVNS) is proposed as a main framework for solving PCTSP. It harnesses a partial optimization metaheuristic under special intensification conditions to prepare candidate sets. Moreover, a topological sort-based greedy algorithm is developed to obtain a feasible solution at the initialization phase. Next, mutation and multi-insertion of constraint-preserving exchanges are combined to produce different neighborhoods of the current solution. Two kinds of constraint-preserving k -exchange are adopted to serve as a strong local search means. Extensive experiments are conducted on 34 cases. For the sake of comparison, Lin-Kernighan heuristic, two genetic algorithms and three VNS methods are adapted to PCTSP and fine-tuned by using an automatic algorithm configurator-irace package. The experimental results show that PVNS outperforms them in terms of both search ability and convergence rate. In addition, the study of four PVNS variants each lacking an important operator reveals that all operators play significant roles in PVNS.

Surrogate-based-constrained optimization for some optimization problems involving computationally expensive objective functions and constraints is still a great challenge in the optimization field. Its difficulties are of two primary types. One is how to handle the constraints, especially, equality constraints; another is how to sample a good point to improve the prediction of the surrogates in the feasible region. Overcoming these difficulties requires a reliable constraint-handling method and an efficient infill-sampling strategy. To perform inequality- and equality-constrained optimization of expensive black-box systems, this work proposes a hybrid surrogate-based-constrained optimization method (HSBCO), and the main innovation is that a new constraint-handling method is proposed to map the feasible region into the origin of the Euclidean subspace. Thus, if the constraint violation of an infeasible solution is large, then it is far from the origin in the Euclidean subspace. Therefore, all constraints of the problem can be transformed into an equivalent equality constraint, and the distance between an infeasible point and the origin in the Euclidean subspace represents the constraint violation of the infeasible solution. Based on the distance, the objective function of the problem can be penalized by a Gaussian penalty function, and the original constrained optimization problem becomes an unconstrained optimization problem. Thus, the feasible solutions of the original minimization problem always have a lower objective function value than any infeasible solution in the penalized objective space. To improve the optimization performance, kriging-based efficient global optimization (EGO) is used to find a locally optimal solution in the first phase of HSBCO, and starting from this locally optimal solution, RBF-model-based global search and local search strategies are introduced to seek global optimal solutions. Such a hybrid optimization strategy can help the optimization process converge to the global optimal solution within a given maximum number of function evaluations, as demonstrated in the experimental results on 23 test problems. The method is shown to achieve the global optimum more closely and efficiently than other leading methods.

As a constrained optimization strategy, model predictive control (MPC) is widely used in the optimal dispatch of microgrids. Rolling optimization and feedback correction strategy can effectively deal with the impact due to the system uncertainty and model mismatch which ensures the stable operation of the microgrid. Due to the uncertainty of distributed power sources, MPC can be applied to solve the problem of renewable energy output in microgrid for fishery. This paper reviews the optimal dispatch of microgrid for fishery based on model predictive control. Firstly, this article introduces the basic principles and classification of model predictive control, and then explains the remaining problems and research directions. The characteristics of the microgrid optimal dispatching problem are described, and the analysis are focused on how MPC solves the uncertain problems in the microgrid optimal dispatching. Finally, perspectives and the future research direction of model predictive control in microgrid energy dispatch is reviewed.

The popularity of deep learning techniques renewed the interest in neural architectures able to process complex structures that can be represented using graphs, inspired by Graph Neural Networks (GNNs). We focus our attention on the originally proposed GNN model of Scarselli et al. 2009, which encodes the state of the nodes of the graph by means of an iterative diffusion procedure that, during the learning stage, must be computed at every epoch, until the fixed point of a learnable state transition function is reached, propagating the information among the neighbouring nodes. We propose a novel approach to learning in GNNs, based on constrained optimization in the Lagrangian framework. Learning both the transition function and the node states is the outcome of a joint process, in which the state convergence procedure is implicitly expressed by a constraint satisfaction mechanism, avoiding iterative epoch-wise procedures and the network unfolding. Our computational structure searches for saddle points of the Lagrangian in the adjoint space composed of weights, nodes state variables and Lagrange multipliers. This process is further enhanced by multiple layers of constraints that accelerate the diffusion process. An experimental analysis shows that the proposed approach compares favourably with popular models on several benchmarks., Published in: IEEE Transactions on Pattern Analysis and Machine Intelligence. arXiv admin note: text overlap with arXiv:2002.07684

This paper proposes a holistic framework of data-driven distributionally robust joint chance constrained economic dispatch (ED) optimization, which seamlessly incorporates deep learning-based optimization for effective utilization of renewable energy in power systems. By leveraging a deep generative adversarial network (GAN), an f-divergence-based ambiguity set of wind power distributions is constructed as a ball centered around the probability distribution induced by a generator neural network. In particular, the GAN is well suited for capturing complicated spatial and temporal correlations of wind power. Based upon this ambiguity set, a distributionally robust joint chance constrained ED model is developed to hedge against distributional uncertainty present in multiple constraints, without assuming a perfectly known probability distribution. The proposed deep learning based ED optimization framework greatly mitigates the conservatism inflicting on distributionally robust individual chance constrained optimization. Theoretical a priori bound on the required number of synthetic wind power data generated by GAN is explicitly derived for the multi-period ED problem to guarantee a predefined risk level. The effectiveness and scalability of the proposed approach are demonstrated in the six-bus and IEEE 118-bus systems by comparing with the state-of-the-art method.

Hyperspectral images containing tens to hundreds of bands can be used in various image classification tasks. However, due to the high data redundancy of the spectral information, the acquiring and analysis of hyperspectral images are usually relatively time-consuming and wasteful of storage space, and therefore limit the practical application of hyperspectral images. Selecting a subset of bands without sacrificing classification accuracy is a strategy to relieve such problems. In this letter, we present an optimization-based method, which can jointly optimize the band selection and the classification network parameters for hyperspectral images. The proposed method regards the discrete selection problem as a continuous constrained optimization problem and adaptively selects the informative band subsets for classification. Besides, experimental results on three public data sets show that our band selection method outperforms the state-of-the-art methods in terms of classification accuracy.

Objectives : Budgetary constraints force healthcare authorities to set priorities for optimal vaccine interventions. A comprehensive decision-making tool would help inform the best combination and sequence of introduction of vaccines within constrained budgets. Methods : Looking at available vaccines against pneumococcal infections in Taiwan (10/13-valent pneumococcal conjugate vaccines [PCV10, PCV13], and 23-valent pneumococcal polysaccharide vaccine [PPV23]), a constrained optimization (CO) model was used to assess the optimal combination of vaccines in children and older adults that would maximize the quality-adjusted life years under pre-defined budget constraints. Scenario analyses were carried out to evaluate the impact of vaccine efficacy (VE) on the optimized solution. Results : The CO model demonstrated that the optimal sequence of vaccine introduction was PPV23 in older adults and PCV10 in children. The optimal solution was mostly driven by the potential to reduce disease burden in the older adult population. VE of PPV23 in older adults and the VE of PCV vaccines against serotype 19A invasive pneumococcal disease had little impact on the optimal solution. Conclusions : The CO approach can be used to set priorities for introducing new vaccines to maximize health gains per age group within the constrained National Vaccine Fund for the prevention of pneumococcal disease in Taiwan.

In this work, we address the problem of making decisions based on data streams, i.e., choosing an action when a new value is recorded. For instance, actions can be trading decisions in financial markets, choices of controllers in dynamic systems or perturbations of the data stream itself. To start with, we propose a language that allows individuals to formulate requirements on the action space. We use prediction techniques to identify the best possible action. However, for many scenarios there is not just one technique that predicts the future precisely, and different techniques behave quite differently. Thus, since there is no technique that dominates all the others, our conclusion is to take multiple predictions generated by different techniques into account. While ensemble techniques aggregating the predictions seem promising, existing techniques have issues, such as unnecessary information losses or the need for a predefined quality measure. Thus, we propose a new ensemble approach that weights predictions techniques according to requirements and solves an optimization problem that derives decisions directly from weighted predictions. We apply our solution to data privacy on data streams. For this setting, the benefits provided by prediction techniques have not been studied yet. In three case studies, we show that our solution consistently achieves better decision-making quality than approaches from related work.

Evolutionary computing is an exciting sub-field of soft computing. Many evolutionary algorithm based on the Darwinian principles of natural selection are developed under the umbrella of EC in the last two decades. EAs provide a set of optimal solutions in single simulation unlike traditional optimization techniques for dealing with large-scale global optimization and search problems. Teaching Learning based Optimization (TLBO) is one of the most recently developed EA. TLBO employs a group of learners or a class of learners to perform global optimization search process. The framework of the TLBO consists of two phases, including the Teacher Phase and Learner Phase. The Teacher Phase’ means learning from the teachers and the Learner Phase means learning through interaction among learners. In this paper, we have developed a hybrid TLBO (HTLBO) with aim at to further improve the exploration and exploitation abilities of the baseline TLBO algorithm. The performance of the proposed HTLBO algorithm examined upon using recently designed benchmark functions for the special session of the CEC2017 problems. The experimental results of the proposed algorithm are better than some well-known evolutionary algorithms in terms of proximity and diversity.

Most conventional optimization approaches are deterministic and based on the derivative information of a problem’s function. On the other hand, nature-inspired and evolution-based algorithms have a stochastic method for finding the optimal solution. They have become a more popular design and optimization tool, with a continually growing development of novel algorithms and new applications. Flexibility, easy implementation, and the capability to avoid local optima are significant advantages of these algorithms. In this study, shapes, and shape perturbation limits of a bracket part, which is used in aviation, have been set using the hypermorph tool. The objective function of the optimization problem is minimizing the volume, and the constraint is maximum von Mises stress on the structure. The grey wolf optimizer (GWO) and the moth-flame Optimizer (MFO) have been selected as nature-inspired evolution-based optimizers.

The virtual power plant (VPP) provides an efficient way to manage a bunch of distributed energy resources (DERs). In order to participate in the system operation and power market as a whole, the estimation of VPP power flexibility is indispensable. However, the temporal coupling constraints make it intractable to aggregate the power flexibility directly. In this paper, we formulate the flexibility evaluation problem as a chance constrained optimization model incorporated the randomness of renewables. To make it tractable, the evaluated flexibility is represented by the combination of a virtual generator and a virtual battery which is decomposed by a robust optimization algorithm. Then, a novel high-dimensional polytope based bound shrinking method is proposed to estimate the parameters of the virtual generator and virtual battery. The numerical test results show the proposed method obtain much more precise results with lower conservativeness than those of existing methods. The disaggregation feasibility of the evaluation results is also verified.

The present study proposes an optimal design methodology of tuned liquid sloshing dampers (TLSDs) installed in wind-excited benchmark tall building. The TLSDs design parameters were determined using proposed optimal design methodology based on nonlinear constraint optimization technique. Top floor peak acceleration was considered as the objective function and maximum value of sloshing depth of liquid as a nonlinear constraint for the design purpose. The classical nonlinear shallow water wave theory was used for the simulation of liquid sloshing, and the basic equations of nonlinear liquid sloshing were solved using the Lax Finite Difference Scheme. The governing equations of motion of combined structure-TLSDs system were expressed as state-space variables and structural response was numerically simulated. The findings were presented and assessed on the basis of a comparison between power spectral density functions, time history analysis, root mean square, peak response, and performance criteria for benchmark building with and without TLSDs. The off-tuning effect of TLSDs due to ± 15 % uncertainty in the building stiffness was also investigated. Further, an attempt has been made to present the control effectiveness of TLSDs in time–frequency domain (wavelet scalogram) using continuous wavelet transformation technique. The optimal TLSDs design based on the proposed methodology was found to be quite effective in serviceability-based design of benchmark tall building under wind loads and can be very useful for the practical design of TLSDs.

Deep learning is a timely research direction in machine learning, where breakthrough progress has been made in both academe and industries, bringing promising results in speech recognition, computer vision, industrial control and automation, etc. The motivation of deep learning is primarily to establish a model to simulate the neural connection structure of the human brain. While dealing with complex tasks, deep learning adopts a number of transformation stages to deliver the in-depth description and interpretation of the data. Deep learning achieves exceptional power and flexibility by learning to represent the task through a nested hierarchy of layers, with more abstract representations formed successively in terms of less abstract ones. One of the key issues of existing deep learning approaches is that the meaningful representations can be learned only when their hyperparameter settings are properly specified beforehand, and general parameters are learned during the training process. Until now, not much research has been dedicated to automatically set the hyperparameters, and accurately find the globally optimal general parameters. However, this problem can be formulated as optimization problems, including discrete optimization, constrained optimization, large-scale global optimization, and multiobjective optimization, by engaging mechanisms of evolutionary computation.

Simulation optimization (SO) techniques show a strong ability to solve large-scale problems. In this article, we concentrate on stochastically constrained SO. There are some challenges to tackle the problem: 1) the objective and constraints have no analytical forms and need to be evaluated via simulation; 2) we should make a tradeoff between exploiting around the best solution and exploring more unknown regions; and 3) both the objective value and feasibility determine the quality of a solution. Motivated by these issues, we propose an adaptive Gaussian process-based search (AGPS) to address stochastically constrained discrete SO problems. AGPS fast constructs the Gaussian process for each performance and then builds a new sampling distribution to adaptively balance exploration and exploitation considering the objective function and stochastic constraints. We show that AGPS converges to the set of globally optimal solutions with probability one. Numerical experiments demonstrate the superiority of our method compared with other advanced approaches. Note to Practitioners —Simulation is widely used to model complex and large-scale systems, such as healthcare, transportation, and supply chain logistics. When optimizing these systems, practitioners always assess overall performance by multiple indicators. Inspired by this issue, this article focuses on a general problem that aims to optimize the system’s primary performance while keeping the secondary performance within limits. We propose an adaptive Gaussian process-based method called AGPS to search for a high-quality solution. The bright spot of AGPS is that it can intelligently determine the quality of a solution considering all the stochastic performance and adaptively search the solution space. The merit lightens the burden of practitioners to design specific parameters for different performance indicators. Numerical experiments demonstrate that AGPS can apply to the large-scale discrete optimization problems with smooth function values and shows higher efficiency than existing methods.

Prescriptive analytics is a mechanism that provides the best set of actions to be taken to prevent undesirable outcomes for a given instance. However, this mechanism is prone to privacy breaches if an adversary with subsidiary data is allowed multiple query access to it. So, we propose a differential privacy mechanism in prescriptive analytics to preserve data privacy. Differential privacy can be achieved with the help of sensitivity of the given actions. Roughly speaking, sensitivity is the maximum change in the given set of actions with respect to the change in the given instances. However, a general analytical form for the sensitivity of the prescriptive analytics mechanism is difficult to derive. So, we formulate a nested constrained optimization to solve the problem. We use synthetic data in the experiments to validate the behavior of the differential privacy mechanism with respect to different privacy parameter settings. The experiments with two real-world datasets—Student Academic Performance and Reddit dataset, demonstrate the usefulness of our proposed method in education and social policy design. We also propose a new evaluation measure called the prescription success rate to further investigate the significance of our proposed method.

Correspondence identification is a critical capability for multi-robot collaborative perception, which allows a group of robots to consistently refer to the same objects in their own fields of view. Correspondence identification is challenging due to the existence of non-covisible objects that cannot be observed by all robots, and due to uncertainty in robot perception. In this paper, we introduce a novel principled approach that formulates correspondence identification as a graph matching problem under the mathematical framework of regularized constrained optimization. We develop a regularization term to explicitly address perception uncertainties by penalizing the object correspondences with a high uncertainty. We also introduce a second regularization term to explicitly address non-covisible objects by penalizing the correspondences built by the non-covisible objects. Our approach is evaluated in robotic simulations and real physical robots. Experimental results show that our method is able to address correspondence identification under uncertainty and non-covisibility, and achieves the state-of-the-art performance.

Background Vaccine allocation is a national concern especially for countries such as the Philippines that have limited resources in acquiring COVID-19 vaccines. As such, certain groups are suggested to be prioritized for vaccination to protect the most vulnerable before vaccinating others. Objective The study aims to determine an optimal and equitable allocation of COVID-19 vaccines in the Philippines that will minimize the projected number of additional COVID-19 deaths while satisfying the priority groups for immediate vaccination. Methods In this study, a linear programming model is formulated to determine an allocation of vaccines such that COVID-19 deaths are minimized while the prioritization framework set by the government is satisfied. Data used were collected up to November 2020. Total vaccine supply, vaccine effectiveness, vaccine cost, and projected deaths are analyzed. Results of the model are also compared to other allocation approaches. Results Results of the model show that a vaccine coverage of around 60–70% of the population can be enough for a community with limited supplies, and an increase in vaccine supply is beneficial if the initial coverage is less than the specified target range. Additionally, among the vaccines considered in the study, the one with 89.9% effectiveness and a 183 Philippine peso price per dose projected the lowest number of deaths. Compared with other model variations and common allocation approaches, the model has achieved both an optimal and equitable allocation. Conclusions Having a 100% coverage for vaccination with a 100% effectiveness rate of vaccine is ideal for all countries. However, some countries have limited resources. Therefore, the results of our study can be used by policymakers to determine an optimal and equitable distribution of COVID-19 vaccines for a country/community. Supplementary Information The online version contains supplementary material available at 10.1007/s40258-021-00667-z.

In this article, optimization of the McPherson suspension mechanism of a real car named Arisan is considered. In this regard, a model based on a real-life suspension system is proposed with the least simplification. This model is built in the ADAMS/View software based on the actual size of the suspension mechanism of Arisan. Moreover, the user-written code of the genetic algorithm in C is added as a plug-in to the ADAMS/View software in a completely innovative way to optimize the suspension system. 16 parameters of the suspension system are selected as design variables to wholly handle its geometry. The value of all design variables is optimally found by GA to minimize the variation of the camber angle as an objective function. Comparison of the obtained optimum suspension by the proposed method with the actual suspension system of Arisan shows a 23.5% improvement in the camber variation angle. It is worth noting that the proposed method does not require a mathematical model of the suspension system that leads to some simplifications such as linearization and non-friction joints. The proposed method can be used for modeling and optimization of other nonlinear dynamical systems such as robotics and building structures.

We propose an intelligent reflecting surface (IRS)-aided secure energy-efficient transmission that minimizes transmit power under constraints on the signal-to-noise ratios at the legitimate user and the eavesdroppers. This is achieved by jointly optimizing the beamforming weights at the transmitter and the phase shift coefficients at the IRS. To overcome the non-convexity of the problem, an alternating optimization algorithm is proposed. More specifically, by iteratively calculating the beamforming weights of a transmitter and the phase-shift coefficients at an IRS, the original optimization problem is separately approximated by a second-order cone programming problem. The results of the proposed algorithm are guaranteed to be the same as the solution of the original problem.

Cloud computing provides a cost-effective deploying environment for hosting and executing workflows as its elasticity, scalability and pay-per-use model. Scientific applications are normally compute- or resource-intensive, and how to run them in the cloud while both meeting QoS of users and guaranteeing the benefits of cloud service providers (CSPs) is still a challenge and depends mainly on workflow scheduling. In this article, we propose a hybrid optimization approach, PSO+LOA, i.e., a combination of particle swarm optimization (PSO) and lion optimization algorithm (LOA) for scheduling workflows in the cloud to minimize the total execution time under budget constraints. The main contributions of our work are: (1) A Euclidean distance (ED) aware particle reposition strategy is defined for two close particles, so as to separate them away from each other, hence enhancing the capability of escaping from local optima. (2) To improve the search and convergence efficiency of original PSO, we modify the velocity update equation by introducing adaptive parameters. (3) Inspired by the multiple-swarm co-evolutionary mechanism of LOA, we integrate PSO with LOA to make a good balance between exploration and exploitation during the whole optimization process. Extensive experiments are conducted over well-known scientific workflows with different sizes and types through WorkflowSim. The experimental results demonstrate that in most cases, PSO+LOA outperforms the existing algorithms in the extent of budget constraint satisfiability, solution quality, i.e., it can generate much better solutions which meet the needs of different budget constraints, especially for large-scale applications, such as the average relative deviation index for PSO+LOA and genetic algorithm (GA) are 0.03% and 0.20%, respectively.

Made available in DSpace on 2021-06-25T11:56:26Z (GMT). No. of bitstreams: 0 Previous issue date: 2021-04-14 Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) Fluid flow in centrifugal compressors is complex and turbulent, making it difficult to achieve a robust equipment design. In this work, a computational fluid dynamics (CFD) analysis is conducted on the periodical domain of a CO2 centrifugal compressor impeller and vaneless diffuser, with a 2.85:1 pressure ratio, using ANSYS CFX. The fluid flow is assumed to be steady state, turbulent, and three dimensional. Since polar angles are not considered in the traditional one-dimensional analysis, eight polar angles on hub and shroud were considered for modeling, sensitivity analysis, and optimization. After numerical verification and validation, a sequential sensitivity analysis (SA) was performed to identify non-influential variables. From the same design of experiment (DoE) used by the Morris qualitative SA, a response surface (RS) was trained to perform a quantitative SA by the smoothing spline ANOVA (SS-ANOVA) method. The Morris method was found to be more conservative than SS-ANOVA, keeping more variables as influent for the analysis. Both methods agreed on influential variables ranking. Low computational effort was required to submit the RS to a constrained optimization procedure using the NSGA-II method. The polytropic efficiency of the optimal centrifugal compressor configuration increased 0.7%, keeping the pressure ratio above 2.85, and the required power and outlet temperature below the base compressor. The impact of the polar angles at trailing edge on output variables is higher than leading edge. The optimal centrifugal compressor found is submitted to different mass flow rates and the overall performance for the optimal angles of the trailing edge and leading edge of the impeller was higher than the base compressor. The strategy adopted herein related to qualitative and quantitative sensitivity analysis coupled with response surface and the constrained optimization was shown to be robust, which can be applied to high-dimensional CFD models to reduce the computational cost with suitable results regarding fluid flow phenomena. Sao Paulo State Univ, Ilha Solteira, SP, Brazil Univ Sao Paulo, Sch Engn, Sao Paulo, SP, Brazil Ctr Univ FEI, Sao Bernardo Do Campo, SP, Brazil Fed Univ ABC, Santo Andre, SP, Brazil Sao Paulo State Univ, Ilha Solteira, SP, Brazil CNPq: 306364/2020-4

This article proposes a two-level network-constrained peer-to-peer (P2P) transactive energy for multi-microgrids (MGs), which guarantees the distribution power network security and allows MGs to trade energy with each other flexibly. At the lower level, a P2P transactive energy is employed for multi-MGs to trade energy with each other. A multi-leader multi-follower (MLMF) Stackelberg game approach is utilized to model the energy trading process among MGs. We prove the existence and the uniqueness of the Stackelberg equilibrium (SE) and provide the closed-form expression for SE. For privacy concerns, we provide several distributed algorithms to obtain SE. At the upper level, the distribution system operator (DSO) reconfigures the distribution network based on the P2P transactive energy trading results by applying the AC optimal power flow considering the distribution network reconfiguration. If there are any network violations, DSO requests trading adjustments at the lower level for network security. We reformulate the DSO operation problem in a mixed-integer second-order cone programming (MISOCP) model and ensure its solution accuracy. Numerical results for a 4-Microgrid system, a modified IEEE 33-bus and 123-bus distribution power system show the effectiveness of the proposed transactive model and its solution technique.

Transit network simulation models are often used for performance and retrospective analysis of urban rail systems, taking advantage of the availability of extensive automated fare collection (AFC) and automated vehicle location (AVL) data. Important inputs to such models, in addition to origin-destination flows, include passenger path choices and train capacity. Train capacity, which has often been overlooked in the literature, is an important input that exhibits a lot of variabilities. The paper proposes a simulation-based optimization (SBO) framework to simultaneously calibrate path choices and train capacity for urban rail systems using AFC and AVL data. The calibration is formulated as an optimization problem with a black-box objective function. Seven algorithms from four branches of SBO solving methods are evaluated. The algorithms are evaluated using an experimental design that includes five scenarios, representing different degrees of path choice randomness and crowding sensitivity. Data from the Hong Kong Mass Transit Railway (MTR) system is used as a case study. The data is used to generate synthetic observations used as "ground truth". The results show that the response surface methods (particularly Constrained Optimization using Response Surfaces) have consistently good performance under all scenarios. The proposed approach drives large-scale simulation applications for monitoring and planning., The paper's author names are wrong in the google scholar

We consider a solar microgrid design and dispatch problem using an adaptive stochastic optimization framework. First, we propose a two-stage mixed-integer model for optimal placement and planning of distributed generation (DG) units and energy storage system units. We incorporate time series modeling into stochastic optimization approach to characterize the solar irradiance uncertainty. In the first stage, design decisions (e.g., location and sizing of DGs) are made and in the second stage, dispatch decisions (e.g., how much to generate, how much to store) are made such that electricity demand is met in a reliable and cost effective way. Chance constraints are employed to control the real load shedding within a predefined probability level. Then, we propose a combined sample average approximation (SAA) and linearization technique to solve this problem more efficiently. The advantage of this approach is that no additional binary variables are introduced while reformulating the chance constraints. Computational time, quality of solution, and load shedding percentage are compared with the traditional SAA. Moreover, we carry out a comprehensive out-of-sample simulation on a real-world case study in the state of Arizona assessing the effectiveness of our approach. Numerical experiments demonstrate that the chance constraints are effective tools for control of load shedding in distributed generation and the proposed approach outperforms traditional methods both in terms of true load shedding percentage and computational time.

In this study, we present a neural network that consists of nodes with heterogeneous sensitivity. Each node in a network is assigned a variable that determines the sensitivity with which it learns to perform a given task. The network is trained via a constrained optimization that maximizes the sparsity of the sensitivity variables while ensuring optimal network performance. As a result, the network learns to perform a given task using only a few sensitive nodes. Insensitive nodes, which are nodes with zero sensitivity, can be removed from a trained network to obtain a computationally efficient network. Removing zero-sensitivity nodes has no effect on the performance of the network because the network has already been trained to perform the task without them. The regularization parameter used to solve the optimization problem was simultaneously found during the training of the networks. To validate our approach, we designed networks with computationally efficient architectures for various tasks such as autoregression, object recognition, facial expression recognition, and object detection using various datasets. In our experiments, the networks designed by our proposed method provided the same or higher performances but with far less computational complexity.

Future wearable and portable devices with multiple transmit antennas operating below 6 GHz are constrained by regulatory limitations on the level of electromagnetic radiation a user can be exposed to, measured using the specific absorption rate (SAR). Signaling designs that are optimized to include SAR constraints can improve the performance of uplink transmission. These signaling schemes could include closed-loop beamforming, closed-loop precoding, and space-time coding, which have all been shown to achieve increased rates when optimized as a function of SAR. Previous research addressed SAR constrained optimization only within a single coherence time block. In this paper, we present transmit policies that dynamically allocate user electromagnetic radiation exposure over time. We propose three exposure allocation methods — optimal, uniform, and asymptotic — in the practical case with causal channel state information (CSI), and an on-off transmission approach for the low SAR-to-noise ratio regime. Our results demonstrate that the performance of SAR-aware transmission can be further improved by exploiting frequency and time diversity.

To enhance the machining efficiency, curved thin-walled parts in the aerospace and automobile industries are often formed to the near-net shapes and then subtractive manufactured to their design shapes. The billets for the manufacturing process may lack of the removal material in some regions due to low forming accuracy, which would increase the rejection rate of these parts. However, there is a chance that some rejected billets can be saved by using partial tolerance ranges to compensate the shape errors of the raw billets. Aiming to reduce the rejection rate and save the production cost, this paper attempts to develop a novel approach for constructing process shape of curved thin-walled part, in which its nominal shape was slightly changed within their tolerance ranges to avoid the undercut during the machining process. Firstly, a registration algorithm was used to find an optimal position between the design shape and the on-machine measurement points. Afterwards, a non-linear constrained optimization model was established to search for the feasible points in the tolerance ranges, in which the profile occupancy, thickness occupancy and non-uniformity of machining allowances were carefully considered. Finally, a new process shape was constructed based on these feasible points, which could be used for adaptive machining of curved thin-walled part. The results of a case study validated the feasibility of the proposed process shape construction approach.

With the increasing penetration of renewable resources such as wind and solar, especially in terms of large-scale integration, the operation and planning of power systems are faced with great risks due to the inherent stochasticity of natural resources. Although this uncertainty can be anticipated, the timing, magnitude, and duration of fluctuations cannot be predicted accurately. In addition, the outputs of renewable power sources are correlated in space and time, and this brings further challenges for predicting the characteristics of their future behavior. To address these issues, this paper describes an unsupervised distribution learning method for renewable scenario forecasts that considers spatiotemporal correlation based on generative adversarial network (GAN), which has been shown to generate realistic time series for stochastic processes. We first utilize an improved GAN to learn unknown data distributions and model the dynamic processes of renewable resources. We then generate a large number of forecasted scenarios using stochastic constrained optimization. For validation, we use power generation data from the National Renewable Energy Laboratory wind and solar integration datasets. The simulation results show that the generated trajectories not only reflect the future power generation dynamics, but also correctly capture the temporal, spatial, and fluctuant characteristics of the real power generation processes. The experimental comparisons verify the superiority of the proposed method and indicate that it can reduce at least 50% of the training iterations of the generative model for scenario forecasts.

One of the most important aspects of visual perception is the inference of 3D shape from a 2D retinal image of the outside world. The existence of several valid mapping functions from object to data makes this inverse problem ill-posed and therefore computationally difficult. In human vision, the retinal image is a 2D projection of the 3D world. The visual system imposes certain constraints on the family of solutions in order to uniquely and efficiently solve this inverse problem. This work specifically focused on the minimization of standard deviations of 3D angles (MSDA) for 3D perception. Our goal was to use a Deep Convolutional Neural Network based on biological principles derived from visual area V4 to achieve 3D reconstruction using constrained minimization of MSDA. We conducted an experiment using novel shapes with human subjects to collect data and test the model. The performance of the network largely agreed with how humans estimated novel 3D shapes. The results show that the constraint of MSDA in 3D shape can be implemented in a neural network and produce human-like results. Additional visual constraints can be added to the network in the future to fully test the theory of visual constraints as a basis of 3D shape perception.

Zeroth-order optimization algorithms are an attractive alternative for stochastic optimization problems, when gradient computations are expensive or when closed-form loss functions are not available. Recently, there has been a surge of activity in utilizing zeroth-order optimization algorithms in myriads of applications including black-box adversarial attacks on machine learning frameworks, reinforcement learning, and simulation-based optimization, to name a few. In addition to utilizing the simplicity of a typical zeroth-order optimization scheme, distributed implementations of zeroth-order schemes so as to exploit data parallelizability are getting significant attention recently. This article presents an overview of recent work in the area of distributed zeroth-order optimization, focusing on constrained optimization settings and algorithms built around the Frank–Wolfe framework. In particular, we review different types of architectures, from master–worker-based decentralized to fully distributed, and describe appropriate zeroth-order projection-free schemes for solving constrained stochastic optimization problems catered to these architectures. We discuss performance issues including convergence rates and dimension dependence. In addition, we also focus on more refined extensions such as by employing variance reduction and describe and quantify convergence rates for a variance-reduced decentralized zeroth-order optimization method inspired by martingale difference sequences. We discuss limitations of zeroth-order optimization frameworks in terms of dimension dependence. Finally, we illustrate the use of distributed zeroth-order algorithms in the context of adversarial attacks on deep learning models.

As more and more services are deployed in the cloud datacenter, network traffic is growing exponentially. Virtual machines (VMs) of a virtual cluster (VC) must be allocated on physical machines (PMs) in the datacenter with a certain topology. Each VM needs some resources to run various services. Apparently, allocating VMs in the datacenter as compactly as possible can reduce traffic consumption and avoid bandwidth-related bottlenecks. However, loose allocation scheme can reduce the loss expectation of VMs due to failure possibilities of PMs and switches, and the availability of the VC is increased thereby. To enhance availability and reduce network bandwidth usage, it is significant to determine the scheme of allocating VMs of the VC. In this paper, we first introduce four typical datacenter architectures with network topologies and corresponding cost matrices, and extend to generality. Then we propose a joint optimization function to measure the risk of VC and the core bandwidth usage with a global availability constraint. Subsequently, an evolutionary algorithm is raised to minimize the value of the constrained optimization function. Finally, the evaluation results show the effectiveness of the proposed approach and performance advancement over the existing approaches.

In confirmatory clinical trials, it has been proposed to use a simple iterative graphical approach to construct and perform intersection hypotheses tests with a weighted Bonferroni-type procedure to control type I errors in the strong sense. Given Phase II study results or other prior knowledge, it is usually of main interest to find the optimal graph that maximizes a certain objective function in a future Phase III study. In this article, we evaluate the performance of two existing derivative-free constrained methods, and further propose a deep learning enhanced optimization framework. Our method numerically approximates the objective function via feedforward neural networks (FNNs) and then performs optimization with available gradient information. It can be constrained so that some features of the testing procedure are held fixed while optimizing over other features. Simulation studies show that our FNN-based approach has a better balance between robustness and time efficiency than some existing derivative-free constrained optimization algorithms. Compared to the traditional stochastic search method, our optimizer has moderate multiplicity adjusted power gain when the number of hypotheses is relatively large. We further apply it to a case study to illustrate how to optimize a multiple testing procedure with respect to a specific study objective., Comment: 26 pages, 4 figures and 1 table

Everyone needs to make decisions under uncertainty and with limited resources, e.g., an investor who is building a stock portfolio subject to an investment budget and a bounded risk tolerance. Doing this with current technology is hard. There is a disconnect between software tools for data management, stochastic predictive modeling (e.g., simulation of future stock prices), and optimization; this leads to cumbersome analytical workflows. Moreover, current methods do not scale. To handle a broad class of uncertainty models, analysts approximate the original stochastic optimization problem by a large deterministic optimization problem that incorporates many "scenarios", i.e., sample realizations of the uncertain data values. For large problems, a huge number of scenarios is required, often causing the solver to fail. We demonstrate sPaQLTooLs, a system for in-database specification and scalable solution of constrained optimization problems. The key ingredients are (i) a database-oriented specification of constrained stochastic optimization problems as "stochastic package queries" (SPQs), (ii) use of a Monte Carlo database to incorporate stochastic predictive models, and (iii) a new SummarySearch algorithm for scalably solving SPQs with approximation guarantees. In this demonstration, the attendees will experience first-hand the difficulty of manually constructing feasible and high-quality portfolios, using real-world stock market data. We will then demonstrate how SummarySearch can easily and efficiently help them find very good portfolios, while being orders of magnitude faster than prior methods.

Visible light communication (VLC) is seen as a supplement for fifth-generation (5G) wireless communication in short-range high data rate communication applications [1] . A reliable VLC system relies on an accurate estimate of the position and orientation of the receiver, which corresponds to the six-dimensional positioning problem mentioned in [2] . In this paper, we investigate the simultaneous position and orientation estimation (SPO) problem using received signal strength (RSS), for a visible light system containing multiple LEDs and multiple photodiodes (PDs) (MLMP). Although in general, the position and orientation of the receiver can be represented by a vector and a rotation matrix, respectively, the constraints imposed by the rotation matrix make the numerical optimization in the estimation process cumbersome, e.g, the commonly used constrained optimization method is often very complex and non-robust. Therefore, in this paper, we design two SPO algorithms using the principle of optimization on manifolds, which alleviates the constraints from the rotation matrix. In addition, we propose an initialization algorithm, based on the direct linear transformation (DLT) principle, to obtain an initial estimate in closed-form for the iterative algorithms. To evaluate the performance of the proposed RSS-based SPO algorithms, we derive the Cramer-Rao bound (CRB). In particular, the orientation error component of the CRB corresponds to the intrinsic CRB or the CRB on manifolds, which measures the error in the estimated rotation matrix in a physically meaningful way. Finally, computer simulations show an asymptotic tightness between the performance of the proposed algorithms and the theoretical lower bound, demonstrating the effectiveness of the proposed solutions.

After the discovery of adversarial examples and their adverse effects on deep learning models, many studies focused on finding more diverse methods to generate these carefully crafted samples. Although empirical results on the effectiveness of adversarial example generation methods against defense mechanisms are discussed in detail in the literature, an in-depth study of the theoretical properties and the perturbation effectiveness of these adversarial attacks has largely been lacking. In this paper, we investigate the objective functions of three popular methods for adversarial example generation: the L-BFGS attack, the Iterative Fast Gradient Sign attack, and Carlini & Wagner's attack (CW). Specifically, we perform a comparative and formal analysis of the loss functions underlying the aforementioned attacks while laying out large-scale experimental results on ImageNet dataset. This analysis exposes (1) the faster optimization speed as well as the constrained optimization space of the cross-entropy loss, (2) the detrimental effects of using the signature of the cross-entropy loss on optimization precision as well as optimization space, and (3) the slow optimization speed of the logit loss in the context of adversariality. Our experiments reveal that the Iterative Fast Gradient Sign attack, which is thought to be fast for generating adversarial examples, is the worst attack in terms of the number of iterations required to create adversarial examples in the setting of equal perturbation. Moreover, our experiments show that the underlying loss function of CW, which is criticized for being substantially slower than other adversarial attacks, is not that much slower than other loss functions. Finally, we analyze how well neural networks can identify adversarial perturbations generated by the attacks under consideration, hereby revisiting the idea of adversarial retraining on ImageNet., Accepted for publication in Pattern Recognition Letters, 2020

Image registration is the process of finding an alignment between two or more images so that their appearances match. It has been widely studied and applied to several fields, including medical imaging and biology, where it is related to morphometrics. In this paper, we present a construction of conformal diffeomorphisms which is based on constrained optimization. We consider a set of different penalty terms that aim to enforce conformality, based on discretizations of the Cauchy–Riemann equations and geometric principles, and demonstrate them experimentally on a variety of images.

Reducing the energy consumption of the servers in a data center via proper job allocation is desirable. Existing advanced job allocation algorithms, based on constrained optimization formulations capturing servers' complex power consumption and thermal dynamics, often scale poorly with the data center size and optimization horizon. This article applies deep reinforcement learning to build an allocation algorithm for long-lasting and compute-intensive jobs that are increasingly seen among today's computation demands. Specifically, a deep Q-network is trained to allocate jobs, aiming to maximize a cumulative reward over long horizons. The training is performed offline using a computational model based on long short-term memory networks that capture the servers' power and thermal dynamics. This offline training approach avoids slow online convergence, low energy efficiency, and potential server overheating during the agent's extensive state-action space exploration if it directly interacts with the physical data center in the usually adopted online learning scheme. At run time, the trained Q-network is forward-propagated with little computation to allocate jobs. Evaluation based on eight months' physical state and job arrival records from a national supercomputing data center hosting 1,152 processors shows that our solution reduces computing power consumption by more than 10 percent and processor temperature by more than 4°C without sacrificing job processing throughput. National Research Foundation (NRF) This research was in part supported by the Nation Research Foundation, Prime Minister's Office, Singapore under its Green Buildings Innovation Cluster (GBIC Award No. NRF2015ENC-GBICRD001-012) and Green Data Centre Research (GDCR Award No. NRF2015ENC-GDCR01001-003), and the AlibabaGroup under its project (RefNo. M4062352).

Despite many attempts to establish material grindability in a milling process, it remains very difficult to predict the milling performance. In this study, impact milling tests were carried out under varying operational conditions in a UPZ100 impact pin mill. The product size distribution and fineness of milled alumina particle are reported and analyzed. A multiscale framework coupling discrete element method (DEM) and population balance model (PBM) is proposed to predict the milling performance of the mill. The impact velocity and impact frequency information from DEM is utilized to inform the mill operation parameters of the PBM model at the process scale, whilst the material grindability parameters are evaluated using constrained optimization to match a milling test. The predictions of the product size distribution show very good agreement with the experimental results. This indicates that the multiscale model is promising for optimizing the design and operation of mills.

Camera calibration is a basic and crucial problem in photogrammetry and computer vision. Although existing calibration techniques exhibit excellent precision and flexibility in classical cases, most of them need from 5 to 10 calibration images. Unfortunately, only a limited number of calibration images and control points can be available in many application fields such as criminal investigation, industrial robot and augmented reality. For these cases, this paper presented a two-step calibration based on soft constraint optimization, which is motivated by "no free lunch" theorem and error analysis. The key steps include (1) homography estimation with weighting function, (2) Initialization based on a simplified model, and (3) soft constraint optimization in terms of reprojection error. The proposed method provides direct access to geometric information of the object from very few images. After extensive experiments, the results demonstrate that the proposed algorithm outperforms Zhang's algorithms from the point of view of the success ratio, accuracy and precision.

Eco-approach and departure (EAD) systems in connected vehicles are capable of providing the speed recommendations to drivers so that vehicles can pass through successive intersections at the appropriate instants to save time and energy. However, most of existing control strategies and global optimization algorithms for EAD systems have extremely high requirements for computational resources, which make EAD systems not easy to popularize and apply to real vehicles. To overcome this problem, this paper designs a traffic flow prediction model based on deep learning regression machine, and establishes a dynamic effective red-light duration model based on traffic flow queuing effect. To facilitate real-time update of the optimal speed, a constrained optimization model is proposed as an approximation approach, which can obtain similar optimal results to that of the pseudo-spectral method while greatly reducing calculating time. The effectiveness of the proposed control algorithm for EAD system on passing intersections and energy saving has been validated on the real vehicle under real-world traffic environment. Compared with the uninformed driver, the proposed EAD system saves 2% time and 4.6% energy on average in case that driver is informed and 8% time and 12.1% energy in case of autonomous driving.

To enhance the flexible interactions among multiple energy carriers, i.e., electricity, thermal power and gas, a coordinated programming method for multi-energy microgrid (MEMG) system is proposed. Various energy requirements for both residential and parking loads are managed simultaneously, i.e., electric and thermal loads for residence, and charging power and gas filling requirements for parking vehicles. The proposed model is formulated as a two-stage joint chance-constrained programming, where the first stage is a day-ahead operation problem that provides the hourly generation, conversion, and storage towards the minimization of operation cost considering the forecasting error of photovoltaic output and load demand. Meanwhile, the second stage is an on-line scheduling which adjusts the energy scheme in hourly time-scale for the uncertainty realizations. Simulations have demonstrated the validity of the proposed method, i.e., collecting the flexibilities of thermal system, gas system, and parking vehicles to facilitate the operation of electrical networks. Sensitivity analysis shows that the proposed scheme can achieve a compromise between the operation efficiency and service quality. Published version

This paper presents a new two-stage multi-swarm particle swarm optimizer (TMPSO), which employs the multi-swarm method and takes two-stage different search strategies in the whole iteration process. This new optimizer includes two versions: unconstrained TMPSO (uTMPSO) and constrained TMPSO (cTMPSO) for unconstrained and constrained global optimizations respectively. For the uTMPSO version, TMPSO makes a certain number of sub-swarms in the first stage to iterate to increase the probability to find the global optimum. Further in the second stage, all the sub-swarms are merged into one large swarm to further refine the global best particle. In both these two stages, each sub-swarm of the first stage and the merged swarm of the second stage all employ a local three-stage multi-point particle swarm optimization (MpPSO) algorithm, which is enlightened by human decision-making and cusp catastrophe theory to enhance the local search ability. To solve constrained optimization problems, the uTMPSO is further upgraded to handle the constraints by using trial and error method to form the cTMPSO version, in which constraints violations are checked on each new created particle in the above uTMPSO procedures, and the violating ones are enforced to execute “retreat” operations, return into the feasible region and recreate new positions, which replaces the traditional penalty function method. This proposed uTMPSO is tested on two unconstrained optimization test functions benchmark set with 25 and 28 functions (including multimodal hybrid composition functions) respectively, and compared with other twelve particle swarm optimization variants. The test results show that uTMPSO has better performance and outperforms most compared algorithms. The cTMPSO is also tested on eight benchmark constrained optimization functions and five engineering application problems.

Automatic summarization is typically aimed at selecting as much information as possible from text documents using a predefined number of words. Extracting complete sentences into a summary is not an optimal way to solve this problem due to redundant information that is contained in some sentences. Removing the redundant information and compiling a summary from compressed sentences should provide a much more accurate result. Major challenges of compressive approaches include the cost of creating large summarization corpora for training the supervised methods, the linguistic quality of compressed sentences, the coverage of the relevant content, and the time complexity of the compression procedure. In this work, we attempt to address these challenges by proposing an unsupervised polynomial-time compressive summarization algorithm. The proposed algorithm iteratively removes redundant parts from original sentences. It uses constituency-based parse trees and hand-crafted rules for generating elementary discourse units (EDUs) from their subtrees (standing for phrases) and selects ones with a sufficient tree gain. We define a parse tree gain as a weighted function of its node weights, which can be computed by any extractive summarization model capable of assigning importance weights to terms. The results of automatic evaluations on a single-document summarization task confirm that the proposed sentence compression procedure helps to avoid redundant information in the generated summaries. Furthermore, the results of human evaluations confirm that the linguistic quality—in terms of readability and coherency—is preserved in the compressed summaries while improving their coverage. However, the same evaluations show that compression in general harms the grammatical correctness of compressed sentences though, in most cases, this effect is not significant for the proposed compression procedure.

The sequencing and line balancing of manual mixed-model assembly lines are challenging tasks due to the complexity and uncertainty of operator activities. The control of cycle time and the sequencing of production can mitigate the losses due to non-optimal line balancing in the case of open-station production where the operators can work ahead of schedule and try to reduce their backlog. The objective of this paper is to provide a cycle time control algorithm that can improve the efficiency of assembly lines in such situations based on a specially mixed sequencing strategy. To handle the uncertainty of activity times, a fuzzy model-based solution has been developed. As the production process is modular, the fuzzy sets represent the uncertainty of the elementary activity times related to the processing of the modules. The optimistic and pessimistic estimates of the completion of activity times extracted from the fuzzy model are incorporated into a model predictive control algorithm to ensure the constrained optimization of the cycle time. The applicability of the proposed method is demonstrated based on a wire-harness manufacturing process with a paced conveyor, but the proposed algorithm can handle continuous conveyors as well. The results confirm that the application of the proposed algorithm is widely applicable in cases where a production line of a supply chain is not well balanced and the activity times are uncertain.

Malicious code has posed a severe threat to modern society. Delivering antivirus program to networks is an important task of a cybersecurity company. As the bandwidth resource in a company is limited and precious, cybersecurity companies have to make a tradeoff between the impact(i.e. the economic loss) of malicious codes and the bandwidth assigned to transmit the antivirus programs. This paper addresses the malicious code and bandwidth tradeoff(MCBT) problem. By developing a novel malicious code and antivirus program interacting model, the total loss, which is the sum of the bandwidth usage fee and the economic loss, is quantified. On this basis, the MCBT problem is modelled as a constrained optimization problem that we refer to as the MCBT model, where the independent variable stands for bandwidth, and the objective function stands for the total loss. Some optimal bandwidth is determined by solving the MCBT model. Based on this, we propose a heuristic algorithm named DOWNHILL, which outperforms random strategies. Finally, the influence of some factors on the optimal bandwidth and the corresponding optimal total loss is uncovered through numerical simulations. To our knowledge, this is the first time the MCBT problem is treated in this way.

The resistance of S-box-based cryptosystems to linear cryptanalysis is often determined by the nonlinearity (NL) and the linear approximation probability (LAP) of the underlying S-box. Constructing dynamic bijective S-boxes with high nonlinearity is a challenging problem. In this paper, we propose a novel S-box construction method based on the concept of constrained optimization. The proposed method uses a random-restart hill-climbing algorithm to construct randomized S-boxes and maximize the nonlinearity of each Boolean function under bijectivity constraints. The proposed algorithm dramatically reduced the S-box construction time. Compared to recent S-box construction methods, the proposed method strikes a better balance among the three design objectives of dynamic S-boxes, namely, cryptographic strength, dynamicity, and speed of construction. On the average, the proposed method constructs a new dynamic $8\times 8$ S-box with NL=112 every 118 ms, whereas a NL=110 S-box can be generated in 5.3 ms, which makes it suitable for real time applications. The proposed method also constructs $8\times 8$ S-boxes with NL=114, which is among the highest reported in literature. Moreover, we demonstrate the extensibility of the proposed constrained optimization formulation to improve other S-box design criteria. Namely, we propose an algorithm to optimize the LAP of an S-box while preserving its NL and bijectivity.

This paper designs the predictive pursuit-evasion game-based orbit control method for the chasing spacecraft to approach a space uncooperative maneuvering target. Firstly, the trajectory planning algorithm combines RRT⁎ and cubic splines to generate a feasible reference trajectory in the relative motion space for the chaser, which enables to deal with the boundary constraints and avoid collisions with the attachments of target. Then, based on the dynamics model, input constrains and objective functions, the pursuit-evasion game model between the target and the chaser is formulated, in which each one acts independently to satisfy its own objective function. With the game formulation, a predictive pursuit-evasion game controller for the chaser is designed to track reference trajectory. Under the frame of model predictive control, the multiple objective constraint optimization can be transferred into to quadratic programming problem to handle input constraints. By predicting opponent’s best move and changing its optimal strategy for its benefit iteratively, the saddle points of the game can be obtained without opponent’s maneuvering. Numerical simulations verify the effectiveness of the predictive pursuit-evasion game control method for approaching an uncooperative target.

In near decades machine learning approaches have received overwhelming attention from many researchers for solving problems that cannot be ironed out by traditional approaches. However, most of these approaches produces output that is not equivalent to the probability estimates of how credible and reliable the output can be for each prediction. One widely utilized, highly accorded for generalized performance but non-probabilistic machine learning algorithm is the Extreme Learning Machine (ELM). As with other classification systems, ELM generates outputs that cannot be treated as probabilities. Current literature shows approaches attempt to assimilate probabilistic concept in ELM however their outputs are not equivalent to probabilities. Furthermore, these methods invoke two-stage post processing procedures with iterative learning procedures which are against the salient features of ELM that highlight no iterative operations involved in learning. Hence, we want to probe in this paper the ability of ELM to produce probabilistic output from the original architecture of ELM itself while preserving the merits of ELM without the need for a post processing two-stage procedures to convert the output to probability and eliminates iterative learning to compute output weights. Two methodologies of unified probabilistic ELM framework are presented, i.e., Probabilistic Output Extreme Learning Machine (PO-ELM) and Constrained Optimization Posterior Probabilistic Outputs based Extreme Learning Machine (CPP-POELM). The proposed models are evaluated empirically on several benchmark datasets as well as real world power system applications to demonstrate its validity and efficacy in handling pattern classification problems as well as decision making process.