Research On The Application Of Firefly Algorithm In Blocking Flow Shop Scheduling Problem

At present, our country’s economy has entered a rapid development stage.Various industries are struggling to international standard. The development ofeconomic brings opportunities. At the same time, it also lets us face unprecedentedchallenges. The21stcentury is an information era of rapid development. As atraditional pillar industry in our country, manufacturing is facing the unprecedentedreform and breakthrough, due to the macroeconomic changes in the world. If thoseenterprises whose main economic income comes from the traditional manufacturingindustry want to get better development, they have to change management pattern,and join more information products to the production process. Besides, they must alsoincrease the productivity and lower the costs for production constantly. Only in thisway, the enterprises could remain competitiveness in the war of global economy.Blocking flow shop scheduling problem, or simply “BFSP” for short, is a veryvital embranchment of flow shop scheduling problem. It has very importanttheoretical significance and application value. In order to enhance the solvingperformance of the blocking flow shop scheduling problem, the firefly algorithm isapplied to the blocking flow shop problem in this paper.In order to apply the firefly algorithm to the discrete scheduling problem, thepaper presents a coding conversion mechanism, which converts real number coding todiscrete encoding skillfully. At the same time, the article also puts forward a doubleinitialization method, and brings in NEH heuristic algorithm in the process ofinitialization. It provides a better search domain for the population, and it’s helpful tofind the global optimal solution. In addition, combined with the levy flight, the articlealso designs a new way to update the individual location, so as to expand the searchdomain, and to keep the firefly algorithm from falling into local optimum. In the end,this paper introduces a kind of local search algorithm. By carrying out the local searchon the individuals of the model with a certain probability, it could enhance the localsearch ability of the firefly algorithm.In the paper, some representative data of Taillard dataset are used as the instances.Particle swarm optimization algorithm, standard firefly algorithm, and the improvedfirefly algorithm are implied to the blocking flow shop scheduling problem. Throughexperimental tests, we conclude that the solving results of the firefly algorithm are more excellent than that of particle swarm optimization. The performance of theimproved firefly algorithm is superior to that of the standard firefly algorithm. Inaddition, the performance of the improved firefly algorithm is more stable, whichshows the effectiveness and robustness of the firefly algorithm in the blocking flowshop scheduling problem.In the article, the firefly algorithm is applied to the blocking flow shopscheduling problem. First of all, the firefly algorithm is improved in many aspects, inorder to improve the performance of the solution, and to strengthen the stability of thealgorithm. To use part of the Taillard data set as the instances, the paper chooses threedifferent intelligent optimization algorithms. Through the tests, we could concludethat the results of improved firefly algorithm have improved greatly, and the solvingperformance is more stable when the improved firefly algorithm is applied to theblocking flow shop scheduling problem. The firefly algorithm has huge potential insolving combinatorial optimization problems.