A Priority-Based Genetic Representations for Bicriteria Network Design Optimizations

Abstract

Network design is one of the most important and most frequently encountered classes of optimization problems. It is a combinatory field in combinatorial optimization and graph theory. A lot of optimization problems in network design arose directly from everyday practice in engineering and management. Furthermore, network design problems are also important for complexity theory, an area in the common intersection of mathematics and theoretical computer science which deals with the analysis of algorithms. Recent advances in evolutionary algorithms (EAs) are interested to solve such practical network problems. However, various network optimization problems typically cannot be solved analytically. Usually we must design the different algorithm for the different type of network optimization problem depending on the characteristics of the problem. In this paper, we investigate the recent related researches, design and validate effective priority-based genetic representations for the typical network models, such as shortest path models (node selection and sequencing), spanning tree models (arc selection) and maximum flow models (arc selection and flow assignment) etc., that these models covering the most features of network optimization problems. Thereby validate that EA approaches can be effectively and widely used in network design optimization.