Knowledge-driven Artificial Bee Colony Algorithm And Its Application For Heterogeneous Flow-shop Scheduling Problem

Flow-shop scheduling problem(FSP)widely exists in various industrial manufacturing and production systems,and FSP belongs to a type of engineering application problem.With the development of intelligent manufacturing,the cooperative production mode between enterprises plays a key role in reducing product costs and improving product quality.Distributed manufacturing has become one of the main manufacturing modes.The no-wait flow-shop scheduling problem is an important branch of the flow-shop scheduling problem.The no-wait constraint of the job requires that the job must be processed continuously without waiting time.The no-wait flow-shop scheduling problem is common in steel manufacturing,food processing,and drug processing.In the actual production system,the difference in factory distribution areas leads to the difference in the production environment of distributed factories.The job of product is assigned to the heterogeneous factory for processing.Jobs are processed through different process routes to obtain the product.Considering the impact of industry heterogeneity,the production scheduling problem in the heterogeneous industry environment is in line with the actual production and processing scenario.Since the scheduling problem is a kind of difficult combinatorial optimization problem,the solution methods of the scheduling problem have experienced the transition from the mathematical method to the artificial intelligence method.The artificial bee colony algorithm(ABC)inspired by the behavior of bees is a swarm intelligence optimization algorithm.Numerous merits of the standard ABC algorithm exist,including simplicity and easy implementation,a small number of parameters,and the robust ability of global exploration.Due to the existence of problem knowledge in practical optimization problems and the existence of empirical knowledge in the iterative process of algorithms,the knowledge-driven artificial bee colony algorithm is designed to address complex optimization problems.The knowledge-driven artificial bee colony algorithm is utilized to address the distributed heterogeneous no-wait flow-shop scheduling problem with sequence-dependent setup times.The main research of the paper is listed as follows.(1)An exploratory landscape analysis driven artificial bee colony algorithm with maximum entropic epistasis(MEEABC)is designed to mitigate defects in the standard ABC algorithm,including the feeble capability of local exploitation and the slower convergence speed.The problem knowledge is extracted from the fitness landscape to guide the search of the algorithm and enhance the local exploration ability of the algorithm.The dimension interaction of the continuous functions computed by the MEE is introduced to guide the search process of the MEEABC algorithm during the employed bee phase and onlooker bee phase.The adaptive mutation methods and the strategy of dynamic population size reduction are implemented to increase the convergence speed and dynamically ameliorate the local exploitation capability.The solutions in the fitness landscape are automatically divided into different clusters to explore the local basin of the fitness landscape via the collaboration between MEE and adaptive mutation methods.Through the effective combination of problem knowledge and algorithm mechanism,the problem knowledge is utilized in the MEEABC algorithm to improve the quality of the population.(2)An improved ABC algorithm combined with multi-agent reinforcement learning(MARLABC)is proposed for addressing the large-scale real value optimization problem in this study.Two stages including the training and the testing are introduced in the MARLABC via the multi-agent central controller to improve the convergence speed and local exploitation capability of the algorithm.The optimal strategy pool is constructed by training procedures via the multi-agent central controller with the Q-learning mechanism.The effective strategy is selected from the optimal strategy pool for each agent during the testing process in the multi-agent central controller.The elite agents in the training population are reserved to generate the testing population to guide the search.The feedback knowledge in the iterative process through training and testing stages is employed in the MARLABC to improve the performance of the algorithm.(3)A distributed heterogeneous no-wait flow-shop scheduling problem with sequence-dependent setup times(DHNWFSP-SDST)is studied in this paper.The differences in factory configuration and transportation time are considered in DHNWFSP-SDST.A mixed-integer linear programming(MILP)model is constructed and an ABC algorithm with Q-learning(QABC)is proposed to address the DHNWFSP-SDST.Heuristic methods named NEH_H and DHHS are designed to construct potential initial candidates for the population.The neighborhood structures based on the job blocks are introduced in QABC to explore the solution space during the evolution processes.The Q-learning mechanism is employed to select neighborhood structures via empirical knowledge in the operation processes.The speed-up methods to accelerate the evaluation of the obtained neighborhood are designed to reduce the computation time of the QABC.The experimental results show that the QABC obtains a high-quality scheduling scheme in a reasonable time.The performance of the MEEABC algorithm and the MARLABC algorithm is tested on the CEC 2017 benchmark test suite.The stability and effectiveness of the MEEABC algorithm and the MARLABC algorithm are confirmed by the experimental results.The experimental results of the MEEABC algorithm and 12 comparison algorithms verify the potential performance of MEEABC in addressing complex optimization problems.The experimental results of the MARLABC algorithm and 14 comparison algorithms show that the MARLABC algorithm is superior to the algorithms in the literature.The performance of QABC for DHNWFSP-SDST is tested on 960 experimental test instances.The experimental results show that the QABC is a potential algorithm to address the DHNWFSP-SDST.