Multi-Agent Differential Evolution Algorithm And Its Applications In Optimization Of Fermentation

Agent which has social, intelligence, adaptability and other characteristics of human beings can simulate human's behavior.Multi-Agent system composed of some agents is more intelligent and has higher abilitiy to solve difficult probleom than a single agent,and it is applied in many fields such as multi-robot systems, intelligent traffic control and etc. Differential Evolution algorithm is a heuristic random search algorithm, which is easy to use, and its stability and powerful capacity of global optimization has been achieved great success in many areas. Study on the combination of the two optimization algorithm and its application problems is of great theoretical and practical significance.In this paper, the research of Multi-Agent system and Differential Evolution are carried out in-depth. Multi-Agent Differential Evolution algorithm is put forward,and it is applied to optimize problems of fermentation process. Major work are as follows:First, Agent and Multi-Agent's basic concepts and ideas are described in detail and summary,and the basic theory of Differential Evolution algorithm is introduced. Based on the respective strengths of Multi-Agent and Differential Evolution algorithm, Multi-Agent's ability of sensing the environment and reacting to the environment and Differential Evolution's capacity of the speed and good global optimization are fully combined,then orthogonal crossover operator and the local optimization operator are bringed forward to form Multi-Agent Differential Evolution algorithm.9 non-binding of 30-1000 dimensional and 9 binding the standard test function have been tested to check the effectiveness of algorithm.Then, Multi-Agent Differential Evolution algorithm is applied to fed-batch optimization problems of alcohol which is growth related fermentation. The simulation results showed that the method is effective to solve this optimization problem and feasibility in the optimization problem of the fermentation process. On this basis, temperature and fed-batch process problems of penicillin fermentation have been studied which is non-growth associated type. And for coordination problem of the process level and loop control level, the whole simulation and optimization method of temperature and fed in penicillin fermentation process is propounded,which take the process level and loop-level to consider. The simulation results show the effectiveness of the method.