Solving Hybrid Flow-shop Schedule By Using Improved Artificial Fish Swarm Algorithm

Hybrid flow shop scheduling problem(Hybrid Flow Shop Scheduling Problem,HFSP), is an extension of traditional flow shop scheduling problem, and its target is toarrange machines and workpieces considering various constraints. Compared with thetraditional flow shop problem, the HFSP with the features of multi-process andmulti-stage is closer to the actual workshop production process. In addition to, there aremany dynamic events in the real world, machine failure is one of the highest frequencydynamic events, and it has a great impact on workshop production process. Therefore,to study the hybrid flow shop scheduling problem has important theoretical significanceand engineering value.It has proved that the hybrid flow shop scheduling problem which even has atwo-stage and only one stage with parallel machines is NP-hard problem, it is difficultto get exact solution. It is very difficult to find a solution to meet the production needsfrom near infinite kinds of scheduling schemes. Artificial fish swarm algorithm as a newbionic optimization algorithm has been widely used in many field, it has the goodfeatures of high adaptability, global optimization and parallel computing, which makesit can be used to solve these scheduling problems efficiently.Therefore, the paper studies how to improve artificial fish swarm algorithm toincrease performance,and use it to solve hybrid flow shop scheduling problem anddynamic hybrid flow shop scheduling problem. The major work is as fellows:First, for the lack of artificial fish swarm algorithm, an improved artificial fishswarm algorithm is proposed, and it introduces parameter adaptive adjustment,sub-regional evolution and communication rules.Secondly, for the hybrid flow shop scheduling problem, the improved artificial fishswarm algorithm has been adjusted. In order to solve the problems of the low number ofthe best individual of the initial population for traditional encoding, and the difficulty ofdecoding for traditional encoding, the another encoding and decoding are designed。Finally, the impact of the relationship between the ready time and repair time to thescheduling performance is discussed. In order to enable the model of HFSP to presentthe hybrid flow shop considering ready time and machine failure, it is be adjusted.Re-scheduling strategy was studied, a combination of two scheduling methods is used tosolve the re-scheduling problem based on the hybrid flow shop characteristics, and binary tree is used to accomplish right shift rescheduling. On this basis, the improvedfish swarm algorithm and rescheduling methods are used to solve the hybrid flow shopscheduling problem under the four different ratio of ready time and repair time. Theresults confirmed the effectiveness of the proposed method.In summary, studying scheduling algorithm, scheduling strategy, and the impact ofready time to scheduling performance with machine failure based on HFSP, in order toprovide a new idea for solving scheduling problems, and make the scheduling theorycan be more practical, and provide a basis for the decision-makers.