Research Of Improved Immune Genetic Algorithms And Its Applications In Optimization Scheduling Problem

With the development of socialization of production and the increase of production scale and circulation of materials as well as the improvement of the complexity, optimization scheduling problem is already impregnated into every domain of scientific research and engineering applications. In modern times, the rapid development of artificial intelligence technology has provided a strong theoretical foundation on solving optimization scheduling problem. Therefore the research in this field has important theoretical meaning and practical value. The traditional optimization scheduling method has many shortcomings which would be unable to effectively apply to large scale complex problems. In recent years, interdisciplinary acrossing research has offered a new train of thought to solve such kind of problem, in which artificial immune optimization algorithm based on the theoretical foundation of imitating biological immune mechanism has demonstrated outstanding performance in various researching and application fields and has become a new hotspot in the fartificial intelligence area.Immune genetic algorithm is subordinate to artificial immune optimization algorithm category. The algorithm put immune thinking into genetic evolution process, which make it have destination to promote population evolving towards optimization trend selectively by using characteristic information. At the same time, it restrains the degraded phenomenon in optimization course. This paper summarizes the theory and characteristics of simple immune genetic algorithm, sums up its shortcomings, improves the algorithm by using various immunology and genetics thinking, and applies the improved algorithm to several typical optimization scheduling problems. Through simulating of examples, the effectiveness and practical value of the algorithm is verified. The main contents and results are as follows:(1) Researching on the artificial immune system and its algorithm deeply, the paper introduces systematacially the biology principle and simulate mechanism of artificial immune system and elaborates the specific content and of artificial immune system. Having analysed the principle and characteristic of simple immune genetic algorithm, the paper emphasizes the lack of simple immune genetic algorithm at stability, convergence and adaptability on solving complex and large scale issues, and puts forward corresponding improving ideas.(2) According to the defects of simple immune genetic algorithm that has the poor local search capacity and premature convergence, the paper draws on the thinking of clonal selection and memory theory in biological immune system, and puts forward a sort of immune clonal algorithm. The proposed algorithm reforms the defects of of simple immune genetic algorithm for its poor local research ability by using clonal operator, increases the searching scope and maintains the diversity of the population through clonal proliferation and super mutation operators,. The paper applies immune clonal algorithms to solving the capacitated vehicle routing problems and verifies the validity and stability of improved algorithm through calculating the Benchmark problems by different sizes and types.(3) According to the ill phenomenons of simple immune genetic algorithm that is easy to fall into equilibrium state and lose advantage genes, the paper puts forward a kind of multi population diploid immune genetic algorithm. On the one hand, the algorithm makes various groups evolving synchronously and exchanges advantaged individuals carrying hereditary information in population, which breaks current equilibrium state in population to reach a higher equilibrium state; On the other hand, it uses the diploid coding mode to prolong the life of useful genic blocks, improve the local search efficiency of algorithm obviously and maintains the diversity of population, which makes the algorithm jump out of local optimal solution. The paper applies the improved algorithm to solving single item capacitated lot-sizing schedule problem. The simulation of examples shows that multi population diploid immune genetic algorithm not only has well overall and local search ability, but also has a well approaching precision and searching speed.(4) According to the defects of simple immune genetic algorithm that designates crossover probability, mutation probability and vaccine statically which makes the search course to perform slowly and even come to a standstill, the paper puts forward a sort of adaptive immune genetic algorithm. Through adjusting crossover and mutation probability adaptively and generating vaccines dynamically, the algorithm improve the shortcomings of simple immune genetic algorithm for slow convergence and vaccine failure. The paper applies the algorithm to solving flexible job-shop scheduling problem, carries out simulation using different algorithm, comparaes the simulation results of adaptive immune genetic algorithm with genetic algorithm and immune genetic algorithm and has a analysis. The contrastive analysis proves that the improved algorithm has a well convergence and robustness.Through studies and analysis of simulation results above, the paper performances comprehensive summary, conclusion and summarize to the improved immune genetic algorithms. In the aspect of handling complex optimization scheduling problem, the paper has put forward some thoughts of improvement, carried out realization and simulation of the algorithm. The paper has a envisagement to the problem for future research and has a prospect on solving optimization scheduling problem using immune genetic algorithm.