How to measure adaptation complexity in evolvable systems - A new synthetic approach of constructing fitness functions

摘要

School of Automation, Wuhan University of Technology, Wuhan 430070, China,Institute of Systems Science and Engineering, Wuhan University of Technology, Wuhan 430070, China;School of Automation, Wuhan University of Technology, Wuhan 430070, China;Institute of Systems Science and Engineering, Wuhan University of Technology, Wuhan 430070, China,Department of Information Technology & Decision Sciences, Old Dominion University, Norfolk, VA 23529, USA;School of Business and Economics, North Carolina A&T State University, Greensboro, NC274U, USA;Institute of Systems Engineering, Huazhong University of Science and Technology, Wuhan 430074, China;School of Automation, Wuhan University of Technology, Wuhan 430070, China;%How to measure the adaptation complexity effectively is an open issue in natural or artificial systems. In this paper, some essential characteristics of adaptation in evolvable systems and the importance/complexity of constructing multi-objective fitness functions in evolutionary computation are analyzed. Based on the authors' previous work on single-objective normalization, a general method is put forward for multi-objective decision making and optimization with its key idea of decomposing the process of constructing fitness functions into their basic units (classes). Then, the issues of determining the corresponding mathematical models and their parameters as well as the issue of integrating all the fitness functions are discussed. Variable weights/objective synthesis is also briefly discussed. A technique in multi-input-multi-output control systems is illustrated to show the usefulness of the method.