Applying evolutionary algorithm in multi-peak decision problems

A typical multiple value decision problem, partnership selection in supply chain management, has been introduced as an example, which is difficult to solve when only using traditional multi-objective decision support methods. To take the self-organization simulation model as the basic complex system theory to analyze the complex interactive mode inside the decision support problems, firstly, the paper briefly introduces a machine learning based decision learning strategy, which could automatically lead to the differentiation of different decision rules. Then, for the only learning strategy that could hardly discover the best solutions in quite a large problem space, an evolutionary algorithm (EA) is imported and designed to help the decision rule discovery procedure. It is difficult to combine EA into a real decision scenario in enterprise life. This paper puts forward an integration simulation model and corresponding algorithm structures which result in satisfying outcomes