Immune Clone Algorithm In Multi-objective Flow Shop Scheduling

The virtuality of production scheduling is a class of optimal scheduling problem, and is a research direction of operation research. This problem can generally be described as:given the premise of production tasks according to the order, the limited human and material resources allocated to different tasks, so as to meet certain specified targets. Typical scheduling problems include the collection of the finished product; a process operation for each product; the equipment or other resources of the various processes, which must be processed according to certain routes to be processed. The goal is to arrange the processing order and the processing start time, to get the order to meet the constraints, while making some performance indicators are being optimized. Production scheduling problem with multiple constraints, multiple objectives, uncertainties and other characteristics, is a typical NP-hard problem. As the key production management, it is important to study the modeling, optimization and to improve the production efficiency.For multi-objective production scheduling problem, this paper studies the related multi-objective optimization theory. Advance a fitness sharing strategy to avoid simply fitting a single objective problem. Draw on genetic algorithm and the basic principles of immune clone algorithm, and combined with improved production scheduling problems, we successfully applied them to flow shop scheduling problem.The main contribution of this paper is as follows:(1) There is not a unique optimal solution in multi-objective optimization problems, so need to find the Pareto solution set. Traditional optimization techniques can only find a Pareto solution, but with the evolutionary algorithm can find out more Pareto non-inferior solutions. This paper presents a genetic algorithm-based fitness sharing strategy, and successfully applied it to continuous function optimization. Simulation results show the feasibility of the algorithm.(2) Immune clonal algorithm refers to the biological immune system, its principle and mechanism, which makes good results for solving engineering optimization problems. In this paper, the basic principles of immune clonal algorithm and framework are used to solve the multi-objective optimization problems. The proposed cloning strategy of the immune has a good effect with non-inferior solution for the Pareto Elitist. (3) Established the multi-objective flow shop scheduling model based on makespan and total flow time of. For multi-objective flow shop problem, proposed an algorithm based on immune clonal classification of non-inferior solutions and crowding distance calculation strategies. By fitness sharing method, the solution to the problem is assessed. With improved immune clonal strategy, the information of non-inferior solutions are found and used effectively. Through mutation the population diversity and algorithm convergence have been improved. The Simulation results of different problems demonstrate the superiority of the model and proposed algorithms.