Interval Optimal Control Arithmetic And Its Appliance Research In The Glutamic Acid Fermentation Process

Interval optimization algorithm is a kind of optimize method, using interval variables instead of dot variables to optimaize. Interval algorithm is a deterministic global method, so that it can get object function’s bounds effectively, and provide the mathematical results strictly. Besides, variables interval can represent uncertainties in the data, therefore these features make the interval method suitable for solving certain nonlinear and parameter uncertainty in control theory. Using the best possible variables interval to replace the best optimaize dot, can give us more options in the practical problems of the engineering. However the existence of the traditional interval algorithm has many weaknesses so that many people do not like to choose this method.Genetic algorithm uses random search techniques, so it can greatly improve the global search capability, but the randomness makes the algorithm often trapped into local optimum. Now although there are many ways to improve the genetic algorithm, but it is still a non-deterministic algorithm. if the search space is quite large, it is still hard to solve the multimodal optimization problems. Therefore, to overcome the shortcomings of traditional interval algorithm and the genetic algorithm, the interval algorithm’s advantage and the genetic algorithm’s advange are combined to form a new hybrid optimization algorithm. Then the hybrid algorithm is applied to the actual production of industrial process. Specific research paper include the following aspects:First, the basic concepts and the basic idea of interval analysis are introduced in detail, and then it illustrates the basic process of this algorithm in solving a specific example. At last introduced a number of application specific in the field of control theory.Then,according to the basic idea of interval algorithm and steps, it prodived improve a new algorithm on the basic of the genetic algorithm, and10diversiform functions are used to determine its reliability.Finally, it is applied to the glutamic acid fermentation process. It uses neural network to found a black box model. According to the actual requirement, the entire fermentation process is divided into single-object optimization and multi-objective optimization:First, put the acid production rate to be a single object to optimize various operating variables. All variables in every moment are present as a interval. Second, the acid production rate and conversion rate are be considered as multi-objective to optimize, and it will also get optimization of interval of each variable.At the end of the paper, research are summarized, and it gives the improving direction of interval algorithm.