Search

Water environment time series prediction is important to efficient water resource management. Traditional water quality prediction is mainly based on linear models. However, owing to complex conditions of the water environment, there is a lot of noise in the water quality time series, which will seriously affect the accuracy of water quality prediction. In addition, linear models are difficult to deal with the nonlinear relations of data of time series. To address this challenge, this work proposes a hybrid model based on a long short-term memory-based encoder-decoder neural network and a Savitzky-Golay filter. Among them, the filter of Savitzky-Golay can eliminate the potential noise in the time series of water quality, and the long short-term memory can investigate nonlinear characteristics in a complicated water environment. In this way, an integrated model is proposed and effectively obtains statistical characteristics. Realistic data-based experiments prove that its prediction performance is better than its several state-of-the-art peers.

Due to the complex dynamic behavior of a Wastewater Treatment Process (WWTP), the existing soft-sensing models usually fail to efficiently and accurately predict its effluent water quality. Especially when a lot of practical data is provided and we do not know which data-pair is more valuable, WWTP modeling becomes a time-consuming process. The main reason is that the existing soft-sensing models update their parameters at each data-pair in one iteration, while some update operations are meaningless. To address this thorny problem, this paper proposes a Deep Belief Network with Event-triggered Learning (DBN-EL) to improve the efficiency and accuracy of soft-sensing model in WWTP. First, some events are defined according to different running condition during the process of training DBN-based soft-sensing model. The different running condition is dominated by the fluctuation of error-reduction rate. Second, an event-triggered learning strategy is designed to construct DBN-EL, whose parameters are updated only when a positive event is triggered. Thirdly, we present the convergence analysis of DBN-EL based on the optimization in a Markov process. Finally, the effectiveness of DBN-EL is demonstrated on soft-sensing of total phosphorus concentration in a practical WWTP system. In experiment, DBN-EL is compared with nine different models on soft-sensing of WWTP. The experimental results show that the efficiency of DBN-EL is 27.6%–64.9% higher than that of nine competitive models, which indicates that the proposed model is readily available for industrial deployment.

Three-dimensional object reconstruction technology has a wide range of applications such as augment reality, virtual reality, industrial manufacturing and intelligent robotics. Although deep learning-based 3D object reconstruction technology has developed rapidly in recent years, there remain important problems to be solved. One of them is that the resolution of reconstructed 3D models is hard to improve because of the limitation of memory and computational efficiency when deployed on resource-limited devices. In this paper, we propose 3D-RVP to reconstruct a complete and accurate 3D geometry from a single depth view, where R, V and P represent Reconstruction, Voxel and Point, respectively. It is a novel two-stage method that combines a 3D encoder-decoder network with a point prediction network. In the first stage, we propose a 3D encoder-decoder network with residual learning to output coarse prediction results. In the second stage, we propose an iterative subdivision algorithm to predict the labels of adaptively selected points. The proposed method can output high-resolution 3D models by increasing a small number of parameters. Experiments are conducted on widely used benchmarks of a ShapeNet dataset in which four categories of models are selected to test the performance of neural networks. Experimental results show that our proposed method outperforms the state-of-the-arts, and achieves about 2.7 % improvement in terms of the intersection-over-union metric.

Cloud computing providers face several challenges in precisely forecasting large-scale workload and resource time series. Such prediction can help them to achieve intelligent resource allocation for guaranteeing that users’ performance needs are strictly met with no waste of computing, network and storage resources. This work applies a logarithmic operation to reduce the standard deviation before smoothing workload and resource sequences. Then, noise interference and extreme points are removed via a powerful filter. A Min–Max scaler is adopted to standardize the data. An integrated method of deep learning for prediction of time series is designed. It incorporates network models including both bi-directional and grid long short-term memory network to achieve high-quality prediction of workload and resource time series. The experimental comparison demonstrates that the prediction accuracy of the proposed method is better than several widely adopted approaches by using datasets of Google cluster trace.

An inference approach is proposed by formulating reasoning processes as particular evolutions of Petri nets. It can be used to design an intelligent agent that executes tasks in a given environment. First, a symbol Petri net is defined to represent a Boolean variable describing a distinct aspect of an environment. Second, a propositional logic sentence in a conjunctive normal form, which may express some background knowledge or a sequence of percepts made by an agent, is formulated as a linear constraint, called as a semantic constraint. Third, an algorithm is constructed to design monitor places enforcing semantic constraints on symbol Petri nets, and its resultant net is called a knowledge Petri net representing relevant knowledge. Fourth, a reasoning algorithm is presented based on a newly defined transition-firing rule of the knowledge Petri net, and can be used to infer or reveal hidden facts. The proposed inference algorithm is efficient since its time computational complexity is proven to be polynomial with respect to the number of Boolean variables. The wumpus world problem is taken as an example to illustrate and verify it.

Nowadays, multi-core architectures are being more and more used in the design of real-time systems. Those systems react usually to their environment that requires to amend their behaviors by applying reconfiguration scenarios. This paper deals with multi-core reconfigurable real-time systems that should be adapted to their environment under real-time constraints. Meanwhile, their synthesis induces a heavy system code and time overhead due to a huge number of threads. The setting up of those systems involves many stages: i) definition of system functionalities, ii) generating of tasks, iii) placement and scheduling of tasks, and iv) generating system code. Correct transition among these steps has an impact on the final system implementation. Thus the need of a designer’s experience is definitely required. However, many problems related to design decisions can be caused due to the complexity of real-time analysis, scheduling, and placement. Those problems may conduct to infeasible implementations. The proposed approach presents a guidance framework to avoid these problems from specification to code generation. This framework is performed by mixed-integer linear programming. It aims to resolve a task partitioning/scheduling problem while optimizing some metrics. The viability of the proposed framework is illustrated by a case study and performance evaluation.

In the context of automated manufacturing systems (AMSs) with unreliable resources, most existing robust deadlock controllers have high computational complexity or relatively low permissiveness. This work focuses on the deadlock control problem of AMSs with a kind of unreliable resources. Petri nets are used to model the dynamic behaviors of such failure-prone AMSs. First a robust deadlock prevention controller is developed for a large class of AMSs under consideration. Such a robust controller guarantees that the system can process all types of parts continuously through any one of their routes, even if one of unreliable resources fails. Also, this robust controller is proved to be optimal, i.e., maximally permissive, during one resource failure period. Then by using the one-step look-ahead method, we establish a polynomial-complexity robust deadlock avoidance policy (DAP) with the same permissiveness as the obtained robust deadlock prevention controller. That is, such a robust DAP not only has low computational complexity, but also is maximally permissive during one resource failure period.

The classification problems with imbalanced datasets widely exist in real word. An Extreme Learning Machine is found unsuitable for imbalanced classification problems. This work applies a Weighted Extreme Learning Machine (WELM) to handle them. Its two parameters are found to affect its performance greatly. The aim of this work is to apply various intelligent optimization methods to optimize a WELM and compare their performance in imbalanced classification. Experimental results show that WELM with a dandelion algorithm with probability-based mutation can perform better than WELM with improved particle swarm optimization, bat algorithm, genetic algorithm, dandelion algorithm and self-learning dandelion algorithm. In addition, the proposed algorithm is applied to credit card fraud detection. The results show that it can achieve high detection performance.

Digital video watermark is widely used to protect copyright and content authentication. DWT-SVD based method, i.e., discrete wavelet transform (DWT) and singular value decomposition (SVD) based method, is one of the most popular state-of-the-art methods. There are two main criteria to evaluate it, i.e., imperceptibility and robustness. The former is measured via the peak signal to noise ratio (PSNR) and structural similarity (SSIM). They should be as high as possible after the watermark is embedded. Robustness measures how easy to restore the embedded watermark by the owner even if the watermarked video is damaged by an outside attacker. Current studies randomly choose embedded positions for watermark, which hardly achieves its highest imperceptibility and robustness. In this paper, we propose a more imperceptible and robust digital video watermarking method than the existing one. First, we introduce a method to measure video frame distortion in a wavelet transform domain. Second, we model the problem to achieve the minimum video frame distortion as an optimization problem, i.e., choose embedded positions that can maximize peak signal to noise ratio. Finally, we evaluate and compare our method with the state-of-the-art approaches by using real experiments and show its advantages in both imperceptibility and robustness.

This paper deals with the real-time scheduling problem of multi-core systems powered by renewable energy harvested from environment. They handle two types of software tasks which are mapped to cores statically and not allowed to migrate. A task can be periodic which may depend on other tasks’ results, or aperiodic which is added to the system to cope with external interruptions. The uncertainty of energy availability in energy harvesting systems makes real-time scheduling more challenging because energy constraints can be violated to ensure real-time performance. A novel scheduling strategy is proposed to effectively compute deadlines allowing for tasks and messages to meet related constraints. This method consists of two phases, (i) the first one defines different time slots each of which is characterized by energy and frequency parameters to cope with the energy availability issue, and (ii) the second one calculates the deadlines ensuring real-time system feasibility by considering the invocation of aperiodic task execution and task precedence constraints. The originality of the current work compared with related studies is that it deals with multi-core, periodic and aperiodic tasks, dependency, energy harvesting, and real-time aspects simultaneously.

Classification technology plays an important role in machine learning. In the process of classification, the presence of noisy samples in datasets tends to reduce the performance of a classifier. This work proposes a clustering-based Noisy-sample-Removed Under-sampling Scheme (NUS) for imbalanced classification. First, the samples in the minority class are clustered. For each cluster, its center is taken as a spherical center, and the distance of the minority class samples farthest from the cluster center is taken as the radius to form a hypersphere. The Euclidean distance from the center of the cluster to every of the majority samples is calculated to decide if they are in the hypersphere. Then, we propose a NUS-based policy to decide if a majority sample in the hypersphere is a noisy sample. Similarly, the noises samples of the minority class are found. Second, We remove noisy-samples from the majority and minority classes and propose NUS. Finally, logistics regression, Decision Tree, and Random Forest are used in NUS as the base classifiers, respectively and compare with Random Under-Sampling (RUS), EasyEnsemble (EE), and Inverse Random Under-Sampling (IRUS) on 13 public datasets. Results show that our method can improve the classification performance in comparison with its state-of-the art peers.

Sharding technology can be used to improve the scalability of a blockchain system. The suitable placement of transactions into shards can reduce cross-shard ones, so as to improve system throughput and reduce delay. An optimization model for transaction placement to blockchain shards is proposed. Based on time analysis and a transaction flow graph, the fitness is computed to decide which shard is the most suitable one to place a transaction. The experimental results show that the proposed model can effectively reduce cross-shard transactions and reduce transaction delay.

Compared with a traditional manufacturing process, 3D printing has advantages of performance and cost in personalized customization and has been applied in many fields. The problem of 3D model orientation optimization is a crucial one in practice. In this paper, based on the mathematical relationship between model orientation and printing time, surface quality, and supporting area, the model orientation problem is transformed into a multi-objective optimization problem with goal of minimizing printing time, surface quality, and supporting area. Ordinal Optimization (OO) is not only applicable to problems with random factors, but also to solve complex deterministic problems. The model orientation is a complex deterministic problem. We solve it with OO in this paper and use linear weighting to convert the multi-objective optimization problem into single-objective one. Finally, we compare the experimental results of solving 3D model orientation problems solved by OO and Genetic Algorithm (GA). The results show that OO requires less calculation time than GA while achieving comparable performance.

Wire rod and bar rolling is an essential process in a steel production system (SPS). This work considers a lexicographic bi-objective scheduling problem originated from it. As a medium process in SPS, both the impacts of its upstream process, i.e., release time constraints, and the requirements of its downstream process, i.e., due time constraints, on SPS performance, must be considered. The concerned problem aims to minimize the number of tardy tasks as a main objective and total setup time as a secondary one. It is solved via a two-stage method. The first-stage problem is handled with an off-the-shelf optimization software, and the second one with iterated greedy algorithms (IGAs). We compare the performance of IGAs under different strategies and select the best one to solve it. By comparing the experimental results an exact single-objective method’s, this paper shows that the proposed method can solve the problem fast with excellent results in terms of the number of tardy tasks and setup time, i.e., the secondary objective, on average. Its usage can result in significant performance improvement in steel production over the existing practice.

In recent years, the Internet of Things (IoT) has become the key enabler for Smart World System (SWS). In SWS environment, the IoT is empowering wireless devices/objects with abilities of sensation and communication with each other via Internet and providing simplified, more efficient, and enjoyable lifestyle. However, efficient communication between various wireless devices in IoT network is very important, especially in long distant wireless transmission case. Cooperative Medium Access Control (MAC) protocols for wireless networks supporting the IoT is becoming popular to handle this issue. It considers several metrics, e.g., SNR (Signal to Noise Ratio), delay and residual energy. Nowadays, distance metric has appended a new dimension in the relay selection metric list for Cooperative MAC. Howsoever, protocols that have used distance metric can be optimized further in selecting a relay node by considering reliability. Hence, targeting to optimize relay selection for designing a cooperative MAC protocol, we have proposed a Reliability-based Distance-aware Data-rate-adaptive Cooperative MAC for wireless IoT networks. For the assurance of unfailing and cooperative data transmission, this work strengthens the most suitable relay node selection using a cooperative reliability factor. Distance and residual energy are also used to ascertain that the position of optimal relay possess superior rates. Extensive simulation results demonstrate that proposed reliability-based optimal relay selection as well as MAC operation significantly improves the performance over the existing ones.

Over the past decade or so, the deadlock control problem of automated manufacturing systems (AMSs) with unreliable resources received a great deal of research attention. Most existing work to date assumes that the studied AMSs contain a single unreliable resource. Most existing control methods belong to the class of deadlock avoidance ones. This paper focuses on deadlock prevention for AMS with multiple unreliable resources. Our goal is to guarantee the continuous production of those parts not requiring any of the failed resources in case of any resource failure. A Petri net model is developed to characterize the failure and repair of unreliable resources. Based on it, a robust deadlock prevention controller that satisfies the above desired property is proposed. Experimental results indicate its effectiveness and superiority over the state-of-the-art methods.

Train safety monitoring and fault diagnosis are critically to prevent the disastrous accidents caused by train failure. Train safety protection sensor network is a modern equipment capable of autonomously monitoring working condition and actively control emergent faults. It strategically places a group of specialized sensors in the vehicles that can record various vital parameters and real-time prompt calculated error code to train driver or dispatcher. This approach also reduces operation expenses by tracing subsystems' reliability, availability and maintainability. Designing such networks faces a number of challenging tasks, as one needs to address some conflicting requirements for quick diagnosis, collaborative decision making, achieving high precision and reliability. This paper presents a novel online Train Safety Sensor Network (TSSN) architecture installed on train emphasis on safety monitoring and diagnostic issues. TSSN consists of multiple sensor layers that monitor train's electrical and mechanical activities, a train data center and a ground data analysis server. TSSN is modular, flexible and able to meet the required safety specification. The hardware and software structure are characterized by lower development costs compared with other market solutions. The server implements fault diagnosis based on a Fault Tree Analysis method (FTA) for a door system. The results show that the sensor network contributes to improve train safety guarantee function.

To overcome long propagation delays for data exchange between the remote cloud data center and end devices in Mobile Cloud Computing (MCC), Mobile Edge Computing (MEC) is merging to push mobile computing, network control and storage to the network edges. A cloudlet in MEC is a mobility-enhanced small-scale cloud, which contains several MEC servers located in close proximity to mobile devices. The main purpose of a cloudlet is to stably provide services to mobile devices with low latency. When a cloudlet offers hundreds kinds of services to millions of mobile devices, it is critical to balance the loads so as to improve performance. In this paper, we propose a three-layer mobile hybrid hierarchical P2P (MHP2P) model as a cloudlet. MHP2P performs satisfactory service lookup efficiency and system scalability as well as high stability. More importantly, a load balance scheme is provided so that the loads of MHP2P can be well balanced with the increasing of MEC servers and query load. A large number of simulation experiments indicate that the proposed scheme is efficient for enhancing the load balancing performance in MHP2P based MEC systems.

Fog computing uses one or more collaborative end users or near-user edge devices to perform storage, communication, control, configuration, measurement and management functions. It can well solve latency and bandwidth limitation problems encountered by using cloud computing. First, this work discusses and analyzes the architectures of Fog computing, and indicates the related potential security and trust issues. Then, how such issues have been tackled in the existing literature is comprehensively reported. Finally, the open challenges, research trends and future topics of security and trust in Fog computing are discussed.

A Bulletin Board System (BBS) yields user-generated posts, which has enjoyed fast spreading speed. Significant events are often revealed by a post. It may then spread widely, thereby producing large influence in some specific social circles and sometimes the whole society. Hence, evaluating post influence becomes important. It can help Web service providers locate quickly those influential posts, users or communities, place right advertisements, expand an event’s influence, and explode a hot topic’s discussions. Recently, BBS has grown to have some new features, e.g., sociability. The existing studies use an Influence Diffusion Model (IDM) and its expanded versions for the analysis of influence. However, they suffer from such drawbacks as identical treatment of every comment or reply, and complete ignorance of relationships among users, thereby leading to the inaccurate assessment of post influence. To overcome the limitations, inspired by our prior user model for user participation in virtual communities, we propose a behavioral model for user participation in a post and give a Sociability-based Influence Diffusion Probability Model (S-IDPM) by utilizing user relationship and reply-chains to measure the responses of different users and evaluate post influence. Experiments with real data collected from a popular BBS. Our results show that S-IDPM outperforms IDM and its expanded version called Influence Diffusion Probability Model (IDPM). S-IDPM can be helpful to achieve better post influence diffusion evaluation than IDM and IDPM do.

As for a class of automated guided vehicle systems (AGVS) that are partially controllable and with undeterministic behavior, a method is presented to design a maximally permissive controller that avoids any collision among vehicles via labeled Petri nets. It is not sure which area a vehicle enters after it leaves one because of insufficient sensors. Furthermore, a vehicle may uncontrollably enter some areas because of insufficient activators. By modeling the former case by multiple transitions labeled by a same symbol, and the latter case by an uncontrollable transition, an AGVS is represented by a labeled Petri net that is partially controllable. The whole net is partitioned into a set of dangerous regions that are controllable such that the uncontrollable issue is addressed. A method is proposed to compute the consistent marking set for a given observed sequence. By the resultant consistent marking set, the marking of each dangerous region is computed, and can be used to design a control action t o prevent each vehicle from entering each dangerous region where there is one AGV already. As a result, the maximally permissive controller is obtained.

Efficient deadlock controllers are critical in the operation of automated manufacturing systems. This work focuses on a deadlock control problem for flexible assembly systems (FAS). Petri nets are used to model the systems. Through their liveness analysis, it characterizes two kinds of structural objects. Each object can lead to a siphon, and may cause the system to deadlock. Based on such objects, a necessary and sufficient condition about the liveness of Petri net models is obtained. In order to prevent each such object from causing FAS to deadlock, a Petri net controller is designed such that its induced siphon cannot be empty. The conjunction of all these controllers is proved to be capable of ensuring deadlock-free operation of a large class of FAS. The effectiveness of the proposed approach is shown via an FAS example.

Petri nets (PNs) are widely adopted for modeling flexible manufacturing systems (FMSs) since they are an effective tool for analyzing the latter’s dynamic behavior and synthesizing a supervisory controller to make a system deadlock-free. As an important subclass of PNs, the weighted system of simple sequential processes with resources (WS 3 PR) can be used to model many FMSs well. Based on the initial marking and structural properties of a WS 3 PR without using marking enumeration and state equations, this paper presents a method to check its liveness. We first define two classes of transitions for a resource subnet of WS 3 PR, based on which, the relationship between markings and strongly connected resource subnets (SCRSs) is analyzed. Next, functions used to check the liveness of a given WS 3 PR are developed, which take the full advantage of SCRS trees and WS 3 PR structure. It is shown that, by the proposed method, the computational complexity for a subclass of WS 3 PR named k -sharing bounded WS 3 PR is polynomial. Sufficient conditions to check liveness of a WS 3 PR are finally established. Two examples are used to illustrate the results.

Efficient siphon computation plays an important role in deadlock control. This work, based on problem decomposition, develops a new method to compute all strict minimal siphons (SMS) in a class of Petri nets called Systems of Simple Sequential Processes with Resources (S 3 PR). It is proved to be of polynomial complexity with respect to the number of SMSs. Therefore, it is readily applicable to an S 3 PR with a large number of SMSs. Its superiority over the existing methods is validated via experimental results.

Remanufacturing end-of-life (EOL) products is an important approach to yield great economic and environmental benefits. A remanufacturing process usually contains three shops, i.e., disassembly, processing and assembly shops. EOL products are dissembled into multiple components in a disassembly shop. Reusable components are reprocessed in a processing shop, and reassembled into their corresponding products in an assembly shop. To realize an overall optimization, we have to integrate them together when making decisions. In practice, a decision-maker usually has to optimize multiple criteria such as cost-related and service-oriented objectives. Additionally, we cannot accurately acquire the detail of EOL products due to their various usage processes. Therefore, multi-objective and uncertainty need to be considered simultaneously in an integrated disassembly-reprocessing-reassembly scheduling process. This work investigates a stochastic multi-objective integrated disassembly-reprocessing-reassembly scheduling problem to achieve the expected makespan and total tardiness minimization. To handle this problem, this work develops a multi-objective discrete fruit fly optimization algorithm incorporating a stochastic simulation approach. Its search techniques are designed according to this problem’s features from five aspects, i.e., solution representation, heuristic decoding rules, smell-searching, vision-searching, and genetic-searching. Simulation experiments are conducted by adopting twenty-five instances to verify the performance of the proposed approach. Nondominated sorting genetic algorithm II, bi-objective multi-start simulated annealing method, and hybrid multi-objective discrete artificial bee colony are chosen for comparisons. By analyzing the results with three performance metrics, we can find that the proposed approach performs better on all the twenty-five instances than its peers. Specifically, it outperforms them by 6.45%–9.82%, 6.91%–17.64% and 1.19%–2.76% in terms of performance, respectively. The results confirm that the proposed approach can effectively and efficiently tackle the investigated problem.

This paper presents three translations from Petri nets to Modeling, Simulation and Verification Language (MSVL) programs. Each translation is directed by one of the three semantics of Petri nets, namely interleaving, concurrency and max-concurrency. Further, for each translation, an equivalence relation between Petri nets and generated MSVL programs is proved. As a result, the supporting tool MSV for MSVL can be used to verify the properties of Petri nets. Case studies are given to show how to do so with MSV under each semantics.

Based on the systems of simple sequential processes with resources (S3PR) model, the existing methods involve prohibitive computation to synthesize a deadlock prevention controller for automated manufacturing systems (AMS). To reduce the computation, this work studies this problem by using a resource-oriented Petri net (ROPN) model. By revealing the relationship between the bad markings and structural properties of an ROPN, it presents a method such that a deadlock prevention controller can be obtained by simple calculation. By such a controller, for each strongly connected subnet in an ROPN, only one control place is needed such that it is structurally very simple. Furthermore, a condition is given under which a maximally permissive controller can be efficiently obtained, which was never seen before. Examples are used to show the application and performance of the proposed method.

Clustering methods play an important role in data mining and various other applications. This work investigates them based on swarm intelligence. It proposes a new clustering method by combining K-means clustering method and mussels wandering optimization algorithm. A single cluster method is well recognized to achieve limited performance when it is compared with a clustering ensemble (CE) that integrates several single ones. Hence, this work introduces a new CE method called weight-incorporated similarity- based CE. The commonly-used datasets with varying size are used to test the performance of the proposed methods. The simulation results illustrate the validity and performance advantages of the proposed ones over some of their peers.

For a deadlock problem in automated manufacturing systems (AMSs) with unreliable resources, the existing control methods mostly belong to the class of deadlock avoidance. This paper focuses on deadlock prevention for AMSs with an unreliable resource. We use Petri nets to model such AMSs and develop their robust deadlock prevention controller. The controller is designed in three layers. In the first layer, the optimal controller is used to ensure that the system can process all types of parts in the absence of resource failures. The function of the second layer controller is to ensure that, when a fault of the unreliable resource occurs at any reachable state, all parts not requiring the faulty resource can be processed and all resources, that they need, but are held by parts requiring the faulty resource for further processing, can be released in order to maximize the resource utilization. The so-called second-level deadlocks caused by the controllers are prevented by the third layer controller. These three controllers together are shown to satisfy the desired properties and hence, able to ensure the robust deadlock-free operation of AMSs with an unreliable resource.

The Minimal number of Control Places Problem (MCPP), which is formulated to obtain optimal and structurally minimal supervisors, needs extensive computation. The current methods to reduce the computational burden have mainly focused on revision of the original formulation of MCPP. Instead, this paper presents methods to accelerate its solution by eliminating its redundant reachability constraints. The optimization problem scale required for supervisor synthesis is thus drastically reduced. First, a sufficient and necessary condition for a reachability constraint to be redundant is established in the form of an integer linear program (ILP), based on a newly proposed concept called feasible region of supervisors. Then, two kinds of redundancy elimination methods are proposed: an ILP one and a non-ILP one. Most of the redundant reachability constraints can be eliminated by our methods in a short time. The computational time to solve MCPP is greatly reduced after the elimination, especially for large-scale systems. The obtained supervisors are still optimal and structurally minimal. Finally, numerical tests are conducted to show the efficiency and effectiveness of the proposed methods.

This paper presents a preliminary study of economical manufacturing from optimal control perspective. We begin with modeling two manufacturing systems as linear and nonlinear constrained continues-time dynamics, respectively Economical manufacturing, characterized by minimal economical cost, is formulated as optimal control problems with variations on the cost function and system dynamics. The optimal control problems are interesting and challenging to solve due to constraints, nonlinearity, battery components, and consideration of energy price profiles. Rigorous analysis establishes the characteristics of optimal solutions and/or properties of optimal control problems. Methods and techniques to solve these optimal control problems are discussed for further investigation.

Supervisory control reconfiguration can handle the uncertainties including resource failures and task changes in discrete event systems. It was not addressed to exploit the robustness of closed-loop systems to accommodate some uncertainties in the prior studies. Such exploitation can cost-efficiently achieve reconfigurability and flexibility for real systems. This paper presents a robust reconfiguration method based on Petri nets (PNs) and integer programming for supervisory control of resource allocation systems (RASs) subject to varying resource allocation relationships. An allocation relationship is seen as a control specification while the execution processes requiring resources as an uncontrolled plant. First, a robust reconfiguration mechanism is proposed. It includes updating the P-invariant-based supervisor and evolving the state of the closed-loop system. The latter adapts to the control specification changes by the self-regulation of the closed-loop system's state. Next, two novel integer programming models for control reconfiguration are proposed, called a reconfiguration model with acceptability and reconfiguration one with specification correction. Since both models integrate the firability condition of transitions, no additional efforts are required for the state reachability analysis. Finally, a hospital emergency service system is used as an example to illustrate them.

In recent years, there are increasing concerns in energy savings for buildings since they are responsible for a large proportion of energy use. A public room in buildings could hold a number of persons who may prefer dissimilar thermal environment. Furthermore, different areas in such rooms may have different temperatures. Also, facility layout in such a room has effect on the distribution of the people in the room. Thus, it may affect its thermal environment and energy consumption as well. It is meaningful and challenging to effectively operate an air-conditioning control (ACC) system by taking the above mentioned factors into account such that the thermal environment is improved and energy is saved. With the lack of research reports on this issue, this work aims at optimally and dynamically controlling the set-point temperature of an ACC system and designing the facility layout so as to maximize the total thermal satisfaction rate (TSR) as well as energy savings. To do so, a non-linear mathematical programming model is proposed to optimize TSR by determining the set-point of an ACC system and the room layout. Then, a particle swarm optimization (PSO) algorithm is constructed to find an optimal or near optimal solution since it is hard to solve a non-linear mathematical programming problem in a reasonable time. Besides, for further energy saving, two more mathematical programming models are proposed to find a set-point of an ACC system under a given outside temperature and room layout determined by the PSO algorithm. Finally, by using a large library room at Macau University of Science and Technology (MUST) as a case, investigations with a large number of experiments are conducted to collect necessary data. Based on the data, regression analysis is done to predict its indoor temperatures in different areas and TSR at a given temperature. Numerical results show that, by the proposed method, it can improve the thermal satisfaction rate by about 27% and save the daily power cost by about 24.3% in comparison with the currently used manual control method.

The work proposes a synthesis method of supervisors for flexible manufacturing systems modeled by a class of generalized Petri nets. A concept of resource usage ratios (RU-ratios) is first presented to describe the occupation degree of a resource by an operation. Next, an intrinsically live structure characterized by a special numerical relationship between arc-weights and initial markings is investigated from a perspective of RU-ratios. Then, a new kind of supervisors is synthesized on the ground of the generic nature of the intrinsically live structure. Such a supervisor can achieve the purposes of both liveness-enforcement and resource usage ratio-enforcement of the system under consideration. Given a plant, it is easy to determine the topological structure of such a supervisor and the number of monitors is bounded by that of resources used in the plant. In addition, when the configuration of the plant model changes, the supervisor can be reusable through adjusting control parameters only without rearrangement of connections. This makes it easy enough and intuitive to be used by industrial practitioners. Instead of maximal behavioral permissiveness, it pursues a precise usage of shared resources that are limited and valuable. Several examples are used to illustrate the proposed methods.

The deadlock control and liveness-enforcement are mostly implemented based on siphons in most existing work using structural theory. Our previous work has revealed a kind of intrinsically live structures in a class of generalized Petri nets called systems of simple sequential processes with weighted resources allocation (WS 3 PR). The liveness-enforcement is achieved by analyzing the structure including weight information and reconfiguring the initial marking of resource places without considering the initial token counts in idle process places. This work extends it to a more general class than WS 3 PR. Instead of siphons, a new concept of structural objects is defined to carry weight information and reflect the circular waits of resources. The numerical relationship between initial markings of resource places and weights of arcs is investigated to yield an intrinsically live structure-based liveness-enforcing method that requires no external monitors. The work provides an insight into structures of generalized Petri nets and a new revenue to achieve the desired deadlock control. Several examples are used to illustrate the proposed method.

This paper introduces an automatic Web service composition method based on logical inference of Horn clauses and Petri nets. The Web service composition problem is transformed into the logical inference problem of Horn clauses by exploring the dependency relations among services. An algorithm is presented, to determine whether the composite service exists in time that is linear in the size of the knowledge base. It can also return the rule set necessary for the composition. Petri nets are chosen to model the rule set and its structural analysis techniques are used to obtain the composite service.

Most existing prevention methods tackle the deadlock issue arising in flexible manufacturing systems modeled with Petri nets by adding monitors and arcs. Instead, this paper presents a new one based on a characteristic structure of WS^3PR, an extension of System of Simple Sequential Processes with Resources (S^3PR) with weighted arcs. The numerical relationships among weights, and between weights and initial markings are investigated based on simple circuits of resource places, which are the simplest structure of circular wait, rather than siphons. A WS^3PR satisfying a proposed restriction is inherently deadlock-free and live by configuring its initial markings. A set of polynomial algorithms are developed to implement the proposed method. Several examples are used to illustrate them.