Research On Hybrid Flowshop Scheduling Problem Based On Immune Particle Swarm Optimization

Hybrid Flowshop Scheduling Problem belongs to the reality of the field ofproduction scheduling problems, which abstracts a simplified model, belongs toenterprise production management, the control of the core part and is common in theprocess of manufacturing. The enterprise to realize the advanced manufacturing andimprove production efficiency, the key is that make full use of good schedulingsolution, effective optimization technique. The domain of production scheduling isthe scheduling algorithm research and application, which has strong academicmeaning and reality value.Particle swarm optimization algorithm is put forward in recent years based onthe swarm intelligence evolutionary algorithm, has the advantages of easy operation,simple for implementation. In the algorithm development’s process, still has someshortcomings, for the performance of the search efficiency is low prematurelyconvergence, into the local extremum, etc. Based on the above reasons, this articlehas improved the particle swarm optimization algorithm, and applied to the hybridflowshop scheduling problem solving.Firstly, this article proposes immune particle swarm optimization algorithmwhich combines immune algorithm and particle swarm optimization algorithm. Itcombines particle swarm optimization algorithm which is convergence speed andimplementation is simple and the advantage of immune mechanism, based on theconcentration of self-regulation mechanism to ensure the diversity of the particle(antibody), it also gives the principles, processes and the algorithm performanceanalysis. Secondly, combining the above optimization theory and particle (antibody)is easy to trap into local optimum in the optimization progress, this article proposes amethod based on improved immune particle swarm optimization algorithm withdynamic disturbance term (IPSO-DDT), which changes the formula fundamentallyand introduces the immune information processing mechanism. It makes the particle (antibody) has immune, memory features and the evolution of the particle has acertain direction, with less number of iterations to find the optimal solution. Thetypical function test results show that the algorithm convergence speed orconvergence precision than those of past progress. Finally, with hybrid flow shopscheduling problem as the research object. The algorithm’s particle (antibody) codingand decoding reference genetic algorithm matrix coding and the target of the problemis minimize completion time, design the appropriate solution algorithm. Simulationresults show that the algorithm can get better scheduling solution, and verify thealgorithm is efficient and practical operability.