Research On Industrial Production Process Optimization Method Based On Case-Based Reasoning

Process optimization technology can increase the production efficiency on existing devices while meet the energy reduction requirements. And companies can gain significant economic benefits using this technology. The research of the industrial process optimization method has both theoretical significance and practical value.This thesis proposes a method of industrial process optimization based on case-based reasoning. Case base is created and expanded. And the accuracy and the speed of decision-making of CBR are improved. And the way to solving process optimization online based on CBR is realized. The main work is as follows:Firstly, this thesis expounds the principle of case-based reasoning. And aiming at the problems of CBR used in process optimization makes the following improvements:genetic algorithms is used to reduce case base and to optimize the case feature attributes’ weights at the same time, making the reasoning decision more accurate and rapid; ant colony clustering algorithm is used to build two stage indexes of case base, reducing the traversing cases when doing case retrieval. And as a result, improves the speed of searching. The improved CBR technology can make optimization decision more accurately and quickly.Secondly, this thesis proposes a method to create source cases using process optimization model. In this method, use the input variables evaluation strategy presented in this thesis and for each set of input values use particle swarm optimization based on divided comparison strategy to solve the optimization model to obtain constructed source case of optimization process based on experience. And then through a case study strategy to build the case base. By using this method to create source cases, case base is enriched and the quality and accuracy of decision-making by CBR is ensured.Finally, a framework of industrial process optimization method based on case-based reasoning is proposed. The framework includes basic process and all steps of using CBR to make optimization decisions for general industrial processes. On the basis of previous two steps of offline preparation, design all components for online application in detail. Applying this method to pellet ingredients achieves satisfying results.