Symbiotic Organisms Search Algorithm And Its Research For The No-wait Flowshop Scheduling Problem

Driven by advanced technologies and personalized demands,the manufacturing industry is undergoing an inevitable transformation towards intelligence and digitalization.As the cornerstone of smart industry,manufacturing systems are confronted with unprecedented challenges.The flow shop scheduling problem plays an indispensable role in the manufacturing system,and the effectiveness of its scheduling solution directly affects the operation of the manufacturing system,the efficiency of the enterprise and the competitiveness of the product.The no-wait flowshop scheduling problem(NWFSP),which has proven to be an NP hard problem,is an important branch of flowshop scheduling that considers the no-wait constraint of no waiting time between processes and is widely used in manufacturing systems such as smelting,chemical and pharmaceutical.Therefore,the design of efficient optimization solutions for solving the no-wait flowshop scheduling problem is nowadays a key research focus in academia and industry.The learning effect refers to the fact that the actual processing time of a workpiece decreases as workers gain experience and improve their processing skills during a production operation.In traditional scheduling models,the machining time of a workpiece is often set as a constant,and the impact of human factors on productivity is not taken into account.Therefore,designing a scheduling model based on the learning effect and proposing an optimisation algorithm that can effectively solve the model is a hot research direction for many scholars at present.Symbiotic organisms search(SOS)algorithm is a novel meta-heuristic optimisation algorithm proposed by modelling symbiotic relationships between organisms.The algorithm is widely used in complex optimization and engineering design due to its easy implementation,parameter-free and strong performance.In this paper,an in-depth study of the SOS algorithm is conducted.By analysing the operation mechanism and advantages and disadvantages of the SOS algorithm,two forms of improvement strategies for the SOS algorithm,continuous and discrete,are proposed to improve the exploration and development capabilities of the algorithm.The improved SOS algorithm is also successfully applied to two no-wait flowshop scheduling problems.The main research and related works in this paper are as follows:(1)A strengthening teaching and learning based multi-population symbiotic organism search algorithm(STLMSOS)is proposed to address the shortcomings of the symbiotic organism search algorithm in solving complex high-dimensional problems,such as low accuracy,poor diversity and the tendency to fall into local optimality.The algorithm divides the ecosystem into three populations based on the evolutionary characteristics of organisms,and draws on the powerful search capability of the "teaching" phase and the excellent search capability of the "learning" phase of the teaching-learning-based optimization algorithm,so that the mutualism population and commensalism population can evolve in stages,enhancing the solution accuracy and diversity of the algorithm.and diversity.Finally,in order to overcome the problem of premature convergence caused by "convergence" in the later stage of the algorithm,a new interaction premise and inter-group interaction mode based on "survival of the fittest" were proposed by studying the complex relationship between biological migration and species evolution.This enables stagnant populations to break out of local optima through timely inter-group interactions.The superiority of the proposed algorithm,STLMSOS,in solving complex high-dimensional problems was verified through a comprehensive evaluation on the CEC2017 problem set.Then,the algorithm was applied to the no-wait flowshop scheduling problem with the objective of maximizing the completion time to test its practical value.The experimental results showed that the algorithm has certain feasibility in solving NWFSP.(2)In response to the NWFSP problem model’s failure to consider the influence of workers on the production cycle,this paper adds considerations for learning effects based on the NWFSP,resulting in the learning effect-based no-wait flowshop scheduling problem(LENWFSP)model.To improve the accuracy and practicality of the problem model,a location truncation learning function that considers human learning limits and inherent properties of workpieces is used.Furthermore,to effectively solve the scheduling model,a tree structure coding mechanism is proposed,which transforms feasible solutions into tree structures suitable for expressing complex problems,achieving high utilization of scheduling information,high exploration of optimization space,and simplification of interaction processes.Finally,simulation experiments verify the effectiveness of the codec strategy based on the tree structure.(3)A tree structure encoding discrete symbiotic organisms search(TSEDSOS)algorithm is proposed for solving the learning effect-based no-wait flowshop scheduling problem by discretizing the solution using a tree structure coding and decoding strategy and combining it with the evolutionary properties of the SOS algorithm.The TSEDSOS algorithm firstly improves the global merit-seeking capability of the algorithm by retaining the "advantageous bloc" of the job in the mutualism population.To solve the problem of the original SOS algorithm not considering the characteristics of the feasible solution space,the population segmentation strategy is adopted to divide the population into advantageous and disadvantageous groups.In this case,the dominant population uses a commensalism population based on local reinforcement to cope with the "big valley" phenomenon,while the weaker population implements a parasitic strategy based on reverse linkage to purposefully seek out promising individuals.Finally,by analyzing the experimental results of the algorithm on the Taillard benchmark dataset,it is known that the TSEDSOS algorithm outperforms other advanced algorithms in solving LENWFSP.