Research On Optimization Method Of Multi-objective Flexible Job Shop Scheduling Problems

Flexible job shop scheduling problem is the extension of traditional job shopscheduling problems, which better reflects real production process and its flexibility. Tostudy the issue has theoretical and practical significance. In this paper, ant colonyoptimization algorithm and particle swarm optimization algorithm are combined as ahybrid method for flexible job shop scheduling problem of multi-objective optimizationis studied.First of all, The background and recent research of flexible shop schedulingproblem are introduced. The theory of swarm intelligence optimization algorithms andits applications are also introduced, focusing on the basic principle ideas from thealgorithm, mathematical models and algorithms processes of ant colony optimizationalgorithm and particle swarm optimization. All of these above have laid a theoreticalfoundation for the proposed hybrid ant colony-particle swarm optimization algorithm.Flexible job shop scheduling problems include route selection process and processscheduling sub-problems. In this paper, we use the idea of decomposition, that is, theant colony optimization algorithm and particle swarm optimization algorithm to solveeach process route selection and process scheduling problem separately. Whendesigning the ant colony optimization algorithm in this paper, using disjunctive graphmodel approach to design the model representation of optional processing machines.Combining with the characteristics of this problem, design the problem solutionstructure diagram and state transition probability of ants. We use position matrixrepresentation in particle swarm optimization algorithm design. The process priorityvalue in particle vector corresponds to the scheduling order, and on this basis, to designthe particle vector decoding method. Finally, the values of main parameters of thealgorithm are determined by the experimental analysis method.In the multi-objective flexible job shop scheduling problem solving, total process time, total machine load and the key machine load three key objectives are optimized atthe same time. A detailed analysis of the relationship between the three optimizationgoals is given. Objective functions of total machine load and the key machine load areminimized in the machine process route selection phase and total process time objectivefunction is minimized in the process scheduling phase. According to the characteristicsof multi-objective optimization, re-design the local heuristic information updatingmethod in ant colony optimization algorithm and calculation formula of the statetransition probability. Simulation results verify the hybrid ant colony-particle swarmoptimization algorithm.In the end, summarize the research of this paper and point put the direction offuture research prospects.