Global optimization of non-convex nonlinear programs using Line-up Competition Algorithm

A global optimization algorithm of simulating evolutionary process, called Line-up Competition Algorithm (LCA), was recently proposed. In the LCA, all families are independent and parallel during evolution. According to the v alue of their objective function, all families are ranked a line-up and are allocated different search spaces based on their positions in the line-up. The preceding excellent families in the line-up gain less search space, wh ich is favorable for local search, accelerating to find optimal point, whil e the latter worse families gain larger search space, which is helpful for global search. Through the competition of two levels of inside a family and between families, the first family in the line-up is continually replaced by other families, or the value of objective function of the first family i s updated continually. As a result, the optimal solution is approached rapi dly. In this paper, the superior performances of the LCA were demonstrated in detail by solving some difficult non-convex nonlinear programming proble ms constrained and unconstrained. (C) 2001 Elsevier Science Ltd. All rights reserved.