| Summary: |
EAs have been applied to a large number of academic and real world problems over the last twenty years or so. In this project, other than Genetic Algorithms (GAs), which are the most popular EAs, we also want to explore other types of EAs, such as Ant Colony Optimization (ACO), Particle Swarm Optimization (PSO), and ElectroMagnetism (EM). Despite some recent implementations, there have been limited applications of these techniques. Nevertheless, there is no evidence that these methods are any less suited for solving combinatorial problems than are genetic algorithms; thus in our opinion there is an urgent need to investigate them further. Our may aim is to build, and analyze, mathematical models for a wide variety of Management Science problems and then implement them computationally in order to obtain solutions. Since the problems we address are NP-hard, the solutions we seek are not necessarily optimal. Therefore, we propose to develop hybrid evolutionary algorithms to obtain such solutions, benefiting from the team previous experience on GAs and hybrid GAs, as well as previous experience on several application fields. Furthermore, advantage can also be taken from the experience with other type of algorithms and from studies made on problem characteristics. In order to fulfill our main objective we have the following specific objectives: 1) study problem characteristics and respective optimal and/or good solutions in order to propose simple problem specific rules or heuristics, mainly to be embedded into EAs; 2) develop some of the above mentioned EAs; 3) develop hybrid EAs by using the results of 1) and 2); and 4) develop a computational application for applying the 3 types of algorithms. |