Research And Application Of An Improved Chaos Immune Algorithm In Job-Shop Scheduling Problem

With the development of global economic integration of production, the production processes of the enterprise products are constantly broken down. Multi-process modes of production become the mainstream. Different from the traditional assembly line production model, multi-process production model is more flexible in the face of changes in production planning and machine faults. Once production fails, the plan can adjust quickly to increase production efficiency. Solving Job-Shop scheduling problem, is profound meaningful to improve production efficiency and the core competitiveness of enterprises.Artificial intelligence technology is developing quickly. Immunology has become an independent discipline of the biomedical field. Each mode is based on the reasonable content of the former model, combined with the latest experimental results, this paper puts forward more complex perfect mechanism to explain and solve new problems in the process of immunology research. Chaotic system is a seemingly random deterministic system. Chaos variable is random and periodic and cannot be limited by local optimum value. Chaos variable can be used to prevent prematurity phenomenon and improve the speed of convergence of the algorithm.In this paper, artificial immune algorithm and chaos theory are combined, an improved chaos immune algorithm is designed to solve Job-Shop scheduling problem. Based on the danger model theory and introducing chaos variables, a basic framework of chaos immune algorithm is built. Then improve each step of the algorithm, mainly focused on improving the speed of convergence of the algorithm, not be limited by local optimal value and the ability to find the global optimum capacity, so to be an efficiency and smart algorithm for solving Job-Shop scheduling problem.A simulation system based on chaos immune algorithm for Job-Shop scheduling problem is realized, and tested by classic benchmark problem and real production data. The algorithm is proved to have good convergence rate and search ability to solve complex problems.