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.

Artificial Bee Colony (ABC) is a Swarm Intelligence optimization algorithm well-know for its versatility. The selection of decision variables to up-date is purely stochastic, incurring in several issues to the local search capability of the ABC. To address these issues, a self-adaptive decision variable selec-tion mechanism is proposed with the goal of balancing the degree of explorationand exploitation throughout the execution of the algorithm. This selection,named Adaptive Decision Variable Matrix (ADVM) represents both stochasticand deterministic parameter selection in a binary matrix and regulates the ex-tent of how much each selection is employed based on the estimation of thesparsity of the solutions in the search space. Influence of the proposed approachto performance and robustness of the original algorithm is validated by experi-menting on fifteen highly multimodal benchmark optimization problems. ADVMis integrated into the original ABC and variations in order to showcase the flex-ibility of the method. Numerical comparison is made against the ABC and theirvariants, as well as to other population-based algorithms (e.g., Particle SwarmOptimization and Differential Evolution). Results show an improvement of theperformance of the algorithms with the ADVM in the most difficult instances.

Artificial Bee Colony (ABC) is a bee inspired swarm intelligence (SI) algorithm well-known for its versatility and simplicity. In crucial steps of the algorithm, employed and scout bees phase, parameters (decision variables) are chosen in a random fashion. Although this randomness may apparently not influence the overall performance of the algorithm, it may contribute to premature convergence towards bad local optima or lack of exploration in multimodal problems featuring rugged surfaces. In this study, a deterministic selection method for decision variables based on Cantor’s proof of uncountability of rational numbers is proposed to be used in the aforementioned steps. The approach seeks to eliminate stochasticity, enhance the exploratory capabilities of the algorithm by verifying all possible variables, and provide a better mechanism to displace solutions out of local optima, introducing more novelty to solutions. In order to analyze potential benefits brought by the proposed approach to the overall performance of the ABC, three variants featuring modifications discussed in this work were designed to be compared in terms of efficiency and stability against the original ABC on 15 instances of unconstrained optimization problems.

The following paper presents a novel strategy named Maximum Search Limitations (MS) for the Evolutionary Particle Swarm Optimization (EPSO). The approach combines EPSO standard search mechanism with a set of rules and position-wise statistics, allowing candidate solutions to carry a more thorough search around the neighborhood of the best particle found in the swarm. The union of both techniques results in an EPSO variant named MS-EPSO. MS-EPSO crucial premise is to enhance the exploration phase while maintaining the exploitation potential of EPSO. Algorithm performance is measured on eight unconstrained and two constrained engineering design optimization problems. Simulations are made and its results are compared against other techniques including the classic Particle Swarm Optimization (PSO). Lastly, results suggest that MS-EPSO can be a rival to other optimization methods.

This article presents a parameter study on the applied game theory in Genetic Algorithm (GA), performing an analysis of the game Prisoner's Dilemma applied in the solution of four constrained Engineering problems. Simulations were applied in four different variations of the game, in order to find the best configuration for the analyzed problems. Then, after obtaining the best configuration of the game, we performed an analysis of the incidence of Alpha and Beta weights, present in GA fitness with Social Interaction, using the game that best suited each problem, employing sixteen different weight configurations. At end, the results obtained authenticated the influence of the Social Interaction process in the essence of GA.

The following paper presents a hybrid Swarm Intelligence algorithm, named Competitive Evolutionary Multi-Colony Optimization (CEMCO). CEMCO is based on the well-known bee inspired algorithm Artificial Bee Colony (ABC), introducing Evolutionary Strategies (ES)under a parallel master-slave architecture utilizing the General Purpose Computing on Graphics Processing Units (GPGPU)paradigm. The goal of this approach is to enhance the overall performance of the master algorithm, producing high quality solutions, and significant solutions under a reduced processing time. Performance evaluation of CEMCO is performed by two large scale constrained optimization problems: Design of Pressure Vessel (DPV)and Welded Beam Design (WBD). Results obtained are compared against prominent techniques found in the literature.

The Artificial Bee Colony (ABC) is a well-known simple and efficient bee inspired metaheuristic that has been showed to achieve good performance on real valued optimization problems. Inspired by such, a Quick Artificial Bee Colony (QABC) was proposed by Karaboga to enhance the global search and bring better analogy to the dynamic of bees. To improve its local search capabilities, a modified version of it, called Evolutionary Quick Artificial Bee Colony (EQABC), is proposed. The novel algorithm employs the mutation operators found in Evolutionary Strategies (ES) that was applied in ABC from Evolutionary Particle Swarm Optimization (EPSO). In order to test the performance of the new algorithm, it was applied in four large-scale constrained optimization structural engineering problems. The results obtained by EQABC are compared to original ABC, QABC, and ABC + ES, one of the algorithms inspired for the development of EQABC.

The Artificial Bee Colony is a popular simple and efficient bee inspired metaheuristic that showed good performance on real valued optimization problems. To improve its local search capabilities, a modified version of it, called ABC+ES was proposed. The new algorithm employs reproduction and mutation operators found in Evolutionary Strategies as well as a reference to the global best from the Particle Swarm Optimization, and has obtained better results in the same field of problems than the Bee Colony. However, testing was only made with problems that have a relatively small number of decision variables. With the intention of exploring the deep learning field, that features domains of high dimensionality, the following paper introduces an adaptation of the Artificial Bee Colony + Evolutionary Strategies for training neural networks. Performance is measured on six representative benchmarks that range from hundreds to thousands of connection weights against other well-known techniques such as exact methods and metaheuristics, as well as the original artificial bee colony. Results show that the approach performed on par with the artificial bee colony for minor scale problems, while it had a better outcome than every other technique on benchmarks with over 1,000 connection weights and multiple outputs. Further study is necessary to confirm whether the approach can be a viable alternative for training networks with very high dimensionality and various outputs aside from gradient methods.

This paper has the purpose to show game theory (GT) applied to genetic algorithms (GA) as a new type of interaction between individuals of GA. The game theory increases the exploration potential of the genetic algorithm by changing the fitness with social interaction between individuals, avoiding the algorithm to fall in a local optimum. To increase the exploitation potential of this approach, this work will present the adjusted crossover operator and compare results to other crossover methods.

The following paper introduces a hybrid algorithm that combines Artificial Bee Colony Algorithm (ABC) and a model of Evolution Strategies (ES) found in the Evolutionary Particle Swarm Optimization (EPSO), another hybrid metaheuristic. The goal of this approach is to incorporate the effectiveness and simplicity of the ABC with the thorough local search mechanism of the Evolution Strategies in order to devise an algorithm that is able to achieve better optimality in less time than the original ABC applied to function optimization problems. With the intention of assessing this novel algorithm performance and reliability, several unconstrained benchmark functions as well as four large-scale constrained optimization-engineering problems (WBD, DPV, SRD-11 and MWTCS) act as an evaluation environment. The results obtained by the ABC+ES are compared to original ABC and several other optimization techniques.

The following paper introduces a parallel approach to a social variant of the Genetic Algorithm, called Parallel Genetic Algorithm with Social Interaction (PSIGA). The algorithm is based on social games involving game theory, and it is implemented using the OpenMP API, which is based on the shared memory programming model for multiple processor architectures. The main contribution of this approach is the parallelization using the Shared Memory of the Social Interaction Genetic Algorithm (SIGA) in order to achieve faster and better optimality than its nonparallel counterpart for global optimization problems with restrictions. For means of performance assessment, the algorithm is tested on four instances of engineering design problems and the obtained results compared with the Genetic Algorithm with Social Interaction (SIGA) implemented in sequential programming model.

This paper has the purpose of presenting a new hybridization of the Artificial Bee Colony Algorithm (ABC) based on the evolutionary strategies (ES) found on the Evolutionary Particle Swarm Optimization (EPSO). The main motivation of this approach is to augment the original ABC in a way that combines the effectiveness and simplicity of the ABC with the robustness and increased exploitation of the Evolution Strategies. The algorithm is intended to be tested on two large-scale engineering design problem and its results compared to other optimization techniques.

This paper presents a new bio-inspired algorithm named Competitive Quantum-Behaviour Evolutionary Multi-Swarm Optimization (CQEMSO) based on CUDA parallel architecture applied to solve engineering problems, using the concept of master/slave swarm working under a competitive scheme and being executed over the paradigm of General Purpose Computing on Graphics Processing Units (GPGPU). The efforts on implementing the CQEMSO algorithm are focused at generating a solution which includes greater quality of search and higher speed of convergence by using mechanisms of evolutionary strategies with the procedures of search and optimization found in the classic QPSO. For performance analysis, the proposed solution was submitted to some well-known engineering problems (WBD, DPV) and its results compared to other solutions found on scientific literature.

In this paper, in order to prove the effectiveness of the Imperialist Competitive Algorithm — a socio-political inspired algorithm-on finding the optimal solution for different kinds of minimization functions as well as different kinds of landscapes. The reliability and quality of solutions for mathematical minimization functions of the ICA is evaluated by seven distinct benchmark functions where each one displays different behaviors, and then the results of each test is compared with two other optimization techniques, the Particle Swarm Optimization (PSO) and the Differential Evolution (DE).