Research On Algorithms For Solving The Distributed Assembly Blocking Flow Shop Scheduling Problems

The efficiency of the distributed flow shop scheduling is the core to the efficient operation and rapid development of intelligent manufacturing systems and advanced manufacturing.The distributed production mode of multi-regional cooperation and multi-factory cooperation has gradually stepped onto the stage of history and has become the main international production mode,which promoted the development of economic globalization and production internationalization.Compared with the traditional single-factory production and processing model,the distributed manufacturing system makes full use of the resources of multiple manufacturing companies or factories,and achieves the distribution of resources by realizing the effective distribution of raw materials,the optimal combination of productivity,and scientific and reasonable resource sharing.The production and manufacturing goals of products are quickly achieved with a reasonable cost in a modern factory.The large-scale manufacturing enterprises adopt intelligent distributed manufacturing mode to maximize production efficiency.Therefore,distributed scheduling plays a vital role in distributed manufacturing systems,and its scientific optimization directly affects the benefits and long-term development of production enterprises.The efficient optimization scheduling methods and technologies is the key to improving production efficiency,saving energy,reducing emissions,reducing production costs,and solving the "stuck neck" problems.Distributed production scheduling problem has the characteristics of large-scale,non-linear,strong constraint,multi-objective,uncertainty,and it is also a hot issue that has always been concerned in the field of optimization and manufacturing.In the context of distributed manufacturing,the distributed production scheduling problem considers the collaborative production of multiple factories,production workshops or production lines.The work distribution of multiple distributed production plants and the processing scheduling of each production line are also studied to achieve the optimization of production indicators.In other words,the distributed production scheduling problem is an inevitable trend of production internationalization under the background of globalization.Firstly,the Monarch Butterfly Optimization algorithm(MBO)is an efficient meta-heuristic to address continuous optimization problems directly.In the MBO algorithm,the mutation and crossover operations are substituted by the migration operator and butterfly adjusting operator.The two operators are cooperative for generating the entire population.Owing to that the individual iterations of the MBO algorithm are not self-learning,the mechanism of the cooperative intelligence is a random process.In this study,an algorithm based on the MBO algorithm with the knowledge-driven learning mechanism(KDLMBO)is presented to strengthen the self-learning capacity of the algorithm.The prior knowledge of the KDLMBO algorithm is the information of the neighborhood extracted from the candidate solutions.The learning mechanism consists of the learning migration operator and learning butterfly adjusting operator.Then,the self-learning collective intelligence is realized by the two cooperative operators in the iterative process of the algorithm.The experimental results demonstrated that the efficiency and significance of the proposed KDLMBO algorithm.Secondly,in the context of production internationalization,the Distributed Assembly Blocking Flow Shop Scheduling Problems(DABFSP)has become an important research hotspot in recent years.This paper obtains the problem characteristics of DABFSP to determine the prior knowledge,and designs a knowledge-driven heuristic method KDH algorithm.The quantitative representation of prior knowledge is the driving force for the design of constructive heuristics,including: knowledge 1,ensuring that all workpieces to be processed belonging to the same product are allocated to different processing plants for parallel processing;knowledge 2,ensuring products in the assembly workshop The assembly sequence is the same as the processing sequence of the workpieces to be processed in the factory;based on knowledge 3,the mechanism of the KDH algorithm ensures that all unprocessed workpieces belonging to the same product are processed together in each factory.According to different sorting rules,the knowledge-driven heuristic algorithm(KDH)proposed in this paper has four variants.The four variants of the KDH algorithm were tested on 900 small-scale instances and 810 large-scale instances,and compared Other 16 latest heuristic algorithms used to solve DAFSP.Combining the experiments of 1710 examples,from the statistical analysis results,the KDH algorithm is feasible and effective.Among the four variants,the KDHLL algorithm has the best performance,so the KDHLL algorithm is selected as a heuristic method to solve the problem of distributed assembly blocking flow shop scheduling.Finally,based on the research of the KDH algorithm,this paper proposes a collaborative learning Monarch Butterfly Optimization Algorithm(CLMBO),and the initialization method of the CLMBO algorithm is KDHLL.When solving the distributed assembly blocking flow shop scheduling problem(DABFSP),the optimization goal of CLMBO is to minimize the total assembly completion time.Similarly,the performance of CLMBO was tested on 1710 instances and compared with the other four latest algorithms.Compared with the other comparison algorithms,CLMBO has a significant advantage in solving the distributed assembly blocking flow shop scheduling problems.Therefore,the CLMBO algorithm is an efficient algorithm for solving the distributed assembly blocking flow shop scheduling problems.