Search

Various optimization problems with multiple decision variables and complex constraints, which exist widely in the real world, are difficult to be solved by traditional methods. Brain storm optimization (BSO) algorithm, an advanced swarm intelligence optimization method, has high efficiency and flexibility in solving large-scale problems independent of problem characteristics. The essence of swarm intelligence optimization algorithm is that a population iteratively searches for the optimal solution in the solution space, and the process has randomness and blindness. To enhance the searching ability of BSO and strengthen the theoretical guidance of the algorithm, a brain storm optimization algorithm with feature information knowledge and learning mechanism (FIBSO) is proposed in this paper. In the process of BSO iteration, information interaction exists between individuals of each generation. The new individual is generated from the old individual, and the dominant individual contributes to the new individual. Theoretically, using the knowledge of characteristic information between individuals guides the evolution of the population in the dominant direction. Moreover, three search strategies guided by global and local optimal individuals are presented to balance the global and local search capabilities of the algorithm. The results of FIBSO and several comparison algorithms on the CEC2017 test suite indicate that FIBSO has superior performance to the state-of-the arts algorithms. The FIBSO is introduced to the no-wait flow shop scheduling problem, and the results show that FIBSO has the significant ability to solve practical engineering problems. [ABSTRACT FROM AUTHOR]

Production scheduling plays a pivotal role in smart factories due to the development of intelligent manufacturing. As a typical scheduling problem, the blocking flow-shop scheduling problem (BFSP) has attracted enormous attention from researchers. In this paper, an ensemble discrete water wave optimization algorithm (EDWWO) is proposed with the criterion to minimize the makespan. In the proposed algorithm, a constructive heuristic is presented to suit the needs of initial solutions quality. The constructive heuristic is based on a new dispatching rule combined with the well-known NEH heuristic. The algorithmic characteristics are explored and effective technologies, such as data-driven mechanism in the propagation phase, a block-shifting operator based on the framework of the variable neighborhood search in the breaking phase, and perturbation strategy, are employed to improve the performance of the algorithm. The effectiveness of operators and parameters in EDWWO are analyzed and calibrated based on the design of experiments. To evaluate the algorithmic performance, the well-known benchmark problem is adopted for comparison with five other state-of-the-art algorithms. Meanwhile, the statistical validity of the results is investigated by introducing the Friedman-test and Wilcoxon-test. The statistical results demonstrate the effectiveness of EDWWO for solving the BFSP. [ABSTRACT FROM AUTHOR]

A reward function is learned from the expert examples by inverse reinforcement learning (IRL), which is more reliable than an artificial method. The moth–flame optimization algorithm (MFO), which is based on the navigation mechanism of a moth flying at night, has been extensively employed to address the complex optimization problem. An inverse reinforcement learning framework with the Q-learning mechanism (IRLMFO) is designed to strengthen the performance of the MFO algorithm in a large-scale real-parameter optimization problem. The right strategy is chosen by the Q-learning mechanism, using historical data provided by the relevant approach in the strategy pool, which stores strategies that include diverse functions. The competition mechanism is designed to strengthen the exploitation capability of the IRLMFO algorithm. The performance of the IRLMFO is verified on the benchmark test suite in CEC 2017. Experimental results illustrate that the IRLMFO outperforms state-of-the-art algorithms. • An inverse reinforcement learning framework with the Q-learning mechanism is designed. • The framework is applied in the metaheuristic algorithm. • The comparison of IRLMFO with state-of-the-art algorithms is presented. [ABSTRACT FROM AUTHOR]

This paper designed a multipopulation cooperative coevolutionary framework with a two-stage orthogonal learning (OL) mechanism for the whale optimization algorithm (MCCWOA) to improve the performance of the whale optimization algorithm (WOA). In the framework, a prediction model of the neighborhood structure is established by discovering the guidance information of the following iteration process in the objective space at the first-stage OL. In the second-stage OL, an auxiliary vector pool with various features in the decision space is introduced to guide the candidates falling in the stagnant status to conduct more valuable exploration. According to the domain knowledge of the candidates, the population is divided into the elite population, the intermediate population, and the inferior population. The information of the subpopulations has interacted with the corresponding historical populations in the evolution processes to enhance the ability of cooperative coevolution among individuals. A standard set of comprehensive benchmark cases and three engineering cases are utilized to verify the advantages of the proposed algorithm. The results of the statistical analysis, diversity analysis, and convergence analysis testified that the MCCWOA outperforms the 15 state-of-the-art algorithms regarding efficiency and significance. [ABSTRACT FROM AUTHOR]

The distributed assembly blocking flow-shop scheduling problem (DABFSP), which is a promising area in modern supply chains and manufacturing systems, has attracted great attention from researchers and practitioners. However, minimizing the total tardiness in DABFSP has not captured much attention so far. For solving the DABFSP with the total tardiness criterion, a mixed integer linear programming method is utilized to model the problem, wherein the total tardiness during the production process and assembly process are optimized simultaneously. A constructive heuristic (KBNEH) and a water wave optimization algorithm with problem-specific knowledge (KWWO) are presented. KBNEH is designed by combining a new dispatching rule with an insertion-based improvement procedure to obtain solutions with high quality. In KWWO, effective technologies, such as the re-developed destruction–construction​ operator, four local search methods under the framework of the variable neighborhood search strategy (VNS), the path-relinking method are applied to improve the performance of the algorithm. Comprehensive numerical experiments based on 900 small-scale benchmark instances and 810 large-scale benchmark instances are conducted to evaluate the performance of the presented algorithm. The experimental results obtained by KWWO are 1 to 4 times better than those obtained by the other comparison algorithms, which demonstrate that the effectiveness of KWWO is superior to the compared state-of-the-art algorithms for the considered problem. • Total tardiness minimization for the distributed assembly blocking flow-shop is investigated. • A constructive heuristic is proposed to obtain competitive initial solutions. • An effective water wave optimization algorithm with problem-specific knowledge is proposed. • Experimental results show the effectiveness of the presented methods. [ABSTRACT FROM AUTHOR]

Brain storm optimization algorithm (BSO), which is inspired by brain storm process of human, has been adopted as an efficient optimizer for various complex problems. A reinforcement learning brain storm optimization algorithm (RLBSO) to improve the performance of BSO is proposed in this paper. Four mutation strategies are designed to enhance the search capability of the algorithm in different stages. Elites are adopted as the guidance to ensure the quality of the population. The global best is utilized to guide the search direction of individuals. The cluster centers are employed to improve the exploitation ability of the algorithm. The historical individuals are utilized to the increase the diversity of the population. The Q-learning mechanism is introduced to guide the selection of strategies according to the historical information fed back by the corresponding strategies. A self-learning mechanism, which is based on the evolutionary state of the population and the experience of previous successful individuals, is utilized to determine the update method of individual. The RLBSO algorithm is tested on the CEC 2017 benchmark test suite and a practical engineering problem. The results show that RLBSO has better performance than the other state of the arts algorithms. • A reinforcement learning brain storm optimization algorithm (RLBSO) is proposed. • Four mutation strategies are introduced to increase the ability of global search. • Q-learning is utilized to guide the algorithm to the appropriate evolution strategy. • A self-learning strategy is utilized to determine the update method of the individual. • The RLBSO algorithm is tested on the CEC 2017 benchmark test suite. [ABSTRACT FROM AUTHOR]

• A novel cooperative learning fruit fly optimization algorithm (HGCLFOA) is proposed. • The hierarchical guidance strategy for local search is implemented in the olfactory search. • The inferior solution repairing (ISR) strategy is employed to modify the search direction. • The hybrid GEDA with previous information is applied to guide the evolution in vision search. • A probability selection strategy based on feedback of objective space is introduced. The fruit fly optimization algorithm (FOA) has drawn enormous attention from researchers and practitioners in the computation intelligence domain for the benefits of simple implementation mechanism and few parameters tuning requirement of FOA. However, FOA is hard to adapt directly to address complex continuous problems. A hierarchical guidance strategy assisted fruit fly optimization algorithm with cooperative learning mechanism (HGCLFOA) is proposed in this study. The population is divided into elitist and inferior subpopulations with the fitness of objective function. The population center is re-designed as an elitist subpopulation to maintain the diversity of the population. In the olfaction search stage, the hierarchical guidance strategy is introduced for local search according to the difference of solution qualities to assign inferior individuals to elitist individuals on different levels. Meanwhile, the inferior information is applied by the inferior solutions repairing strategy to deflect the prediction of the elitist subpopulation for preventing HGCLFOA from falling into the local optimum. In the vision search stage, a hybrid Gaussian distribution estimation strategy is adopted to extract the elitist information of previous generations to predict the distribution of potential elitist individuals in the next generation. The exploration and exploitation of the HGCLFOA are balanced by the cooperation between elitist subpopulation and inferior subpopulation. A random walk strategy is activated to assist the elitist solutions to jump out the local optimal. The parameters of the HGCLFOA are calibrated by DOE and ANOVA methods. The experimental results demonstrated that the HGCLFOA outperformed the classical FOA and state-of-arts variants of FOA. [ABSTRACT FROM AUTHOR]

• The proposed algorithm is effective for solving many-objective problems. • Shift-based estimation and sum of objectives can reserve the excellent solutions. • An angle-based selection strategy can improve the diversity of solutions. • Experimental results have been presented by using Wilcoxon's rank-sum test. Due to the curse of dimensionality, the existing evolutionary algorithms have difficulties in balancing convergence and diversity in many-objective problems. To address this shortcoming, this paper proposes an efficient many-objective optimizer named MaOEA-ASS. In the MaOEA-ASS, the angle-based selection strategy is used to obtain solutions with good diversity from the population. In addition, the combination of the shift-based density estimation and the sum of objectives, which uses the iteration information and emphasis the distribution of solutions, is employed to obtain the high-quality solutions approximating the optimal Pareto solutions. The proposed MaOEA-ASS is compared with eight state-of-the-art many-objective optimization algorithms (MaOEAs) on the DTLZ and WFG test suites, and its performance is verified on a practical many-objective problem. The experimental results demonstrate that the proposed MaOEA-ASS has a superior performance over the peer competitors on all considered many-objective problems. [ABSTRACT FROM AUTHOR]

• A two-stage evolution strategy is proposed for solving multi-objective problem. • The convergence and diversity should be balanced in multi-objective optimization. • A local searching method can improve the diversity of solutions in the space. • Experimental results have been presented by using statistical method. The balance of convergence and diversity plays a significant role to the performance of multi-objective evolutionary algorithms (MOEAs). The MOEA/D is a very popular multi-objective optimization algorithm and has been used to solve various real world problems. Like many other algorithms, the MOEA/D also has insufficient ability of convergence and diversity when tackling certain complex multi-objective optimization problems (MOPs). In this paper, a novel algorithm named MOEA/D-TS is proposed for effectively solving MOPs. The new algorithm adopts two stages evolution strategies, the first stage is focused on pushing the solutions into the area of the Pareto front and speeding up its convergence ability, after that, the second stage conducts in the operating solution's diversity and makes the solutions distributed uniformly. The performance of MOEA/D-TS is validated in the ZDT, DTLZ and IMOP problems. Compared with others popular and variants algorithms, the experimental results demonstrate that the proposed algorithm has advantage over other algorithms with regard to the convergence and diversity in most of the tested problems. [ABSTRACT FROM AUTHOR]