K. Deb et al. (Eds.): GECCO 2004, LNCS 3102, pp. 737–747, 2004. © Springer-Verlag Berlin Heidelberg 2004 A Novel Multi-objective Orthogonal Simulated Annealing Algorithm for Solving Multi-objective Optimization Problems with a Large Number of Parameters Li-Sun Shu 1 , Shinn-Jang Ho 1 , Shinn-Ying Ho 2 , Jian-Hung Chen 1 , and Ming-Hao Hung 1 1 Department of Information Engineering and Computer Science Feng China University, Taichung, Taiwan 407, ROC {p860048@knight, syho@, p8800146@knight, p8800043@knight}.fcu.edu.tw 2 National Huwei Institute of Technology, Huwei, Yunlin, Taiwan 632, ROC Department of Automation Engineering sjho@nhit.edu.tw Abstract. In this paper, a novel multi-objective orthogonal simulated annealing algorithm MOOSA using a generalized Pareto-based scale-independent fitness function and multi-objective intelligent generation mechanism (MOIGM) is proposed to efficiently solve multi-objective optimization problems with large parameters. Instead of generate-and-test methods, MOIGM makes use of a sys- tematic reasoning ability of orthogonal experimental design to efficiently search for a set of Pareto solutions. It is shown empirically that MOOSA is comparable to some existing population-based algorithms in solving some multi-objective test functions with a large number of parameters. 1 Introduction Many real-word applications usually involve simultaneous consideration of multiple performance criteria that are often incommensurable and conflict in nature. It is very rare for these applications to have a single solution, but rather a set of alternative so- lutions. These Pareto-optimal solutions are those for which no other solution can be found which improves along a particular objective without detriment to one or more other objectives. Multi-objective evolutionary algorithms (MOEAs) for solving mul- tobjective optimization problems gain significant attention from many researchers in recent years [1]-[8]. These optimizers not only emphasize the convergence speed to the Pareto-optimal solutions, but also the diversity of solutions. Niching techniques, such as fitness sharing and mating restriction, are employed for finding uniformly distributed Pareto-optimal solutions [2]-[3], and elitism is incorporated for improving the convergence speed to the Pareto front [4]. In recent years, many MOEAs employing local search strategies for further im- proving convergence speed have been successively proposed [4]-[7]. Population-