Research On Reusability Of Discrete Simulation Model Based On CBR

The rapid development of simulation tools provides the facility for the modeler and reduces the difficulty of modeling, however there are still a lot of repetitive work in the modeling work. Through the establishment of a similar model for the reuse and redevelopment, it can effectively reduce repetitive modeling to improve the modeling efficiency.This paper focuses on the reusability of discrete Arena model, the main research of this paper abstracted as following:Based on the in-depth study on the key technologies of Arena software and CBR methods, and combined with the characteristics of discrete Arena model, parameterization of the Arena model to describe the case of the Arena. It Use the Microsoft Access to build the Arena model database and then use data tables to reflect the relationship between attributes of the Arena model. And then Designs a three-level query strategy to test the case library and generate a smaller model set to improve the efficiency of similarity degree.Directing at the various attributes and the different types of the attribute values in the Arena models, it uses the nearest neighbor algorithm to compute the similarity of the attribute According to the connection property between the modules in Arena model. The model is abstracted into directed graphs, while using the similarity algorithm directed graph model to calculate the topology similarity of Arena model. After comprehensive these two algorithms, it derives global similarity calculation algorithm of Arena model. At last, select the PSO algorithm to implement, and test the effectiveness of the algorithm in the final.