Research And Application Of Set-interval Swarm Optimization Algorithm

Interval optimization algorithm takes interval number as the variables.Interval number can represent uncertain data,provide mathematical results strictly and give us more selection space in complex industrial process to prevent noise and disturbance form influencing the system.These features make interval optimization algorithm suitable for solving the optimization of industrial processes with disturbances and uncertain parameters.The control circuit of basis automation has a certain tracking accuracy,So it has no practical significance to illimitably reduce width of interval variable.To solve this problem,this paper put forward a set-interval particle swarm optimization to solving the problem.Meanwhile,this paper establishes BP Neural Networks Model for glutamic acid fermentation process,according to the characteristics of it.Then,the decision variables of the model of glutamic acid fermentation process are optimized by above optimization algorithm.Research contents of this paper mainly include the following aspects:Firstly,basic concepts and operational rules of interval number are introduced,inclusion function's basic knowledge is given out,and modal interval which is an extension of classical interval is briefly presented.Meanwhile the traditional interval optimization algorithm-interval bisection and interval particle swarm optimization were analyzed.Secondly,this paper designs the interval-integral approximation method which solved relativity problem of interval calculation,introduces the method which controls the diversity of method,and used the inner and outer approximations of the ranges of real function to determine the range of the minimum fitness function value of the current interval value.Then,basing on the key technologies,this paper constructed the single target set-interval particle swarm optimization.Thirdly,the basic concept of multi-objective optimization and the already existing interval multi-objective optimization are introduced.The measurement method of crowding distance of the individuals and the deletion strategy of superfluous non-inferior solution are improved.At last,the paper establishes a BP neural network model for the glutamic acid fermentation process.Then,the BP neural network was extended to the interval BP neural network.The glutamic acid fermentation process is divided into two stages.So,the two stages can be respectively optimized by the above single-objective optimization and multi-objective optimization to get the highest acid production rate and the optimal interval value of the decision variables.Further research suggestions are given at the end of this paper.