A New Global Optimization Search Algorithm For Process System

A new global optimization search algorithm, Line-up Competition Algorithm(LCA). is proposed.which proceeds from simulating live things evolution in behavior and stresses on the function of specie population competition on evolutionary driving force. There always exist independent and parallel evolutionary families in the course of simulating evolution. Population can evolve rapidly to optimum or near optimum region by the two different competition fashions of the live struggle inside family and the position competition among families.The concept of competition driving force is proposed and its quantitative descriptions are given. Ranged a line-up based on their objective function value, families gain corresponding competition driving force in the light of their positions in the line-up, not only resulting in the dynamic change of the position of every family in line-up, but also reaching the uniformity of local and global search and unity of individual competition and cooperation.The solution of global optimization problems of a set of typical test function with LCA shows excellent search ability. LCA has obvious advantages over Genetic Algorithms and Simulated Annealing in solution quality and search rate. That mixed-integer nonlinear programming problems, including greater scale multiproduct batch process design, are solved with LCA for chemical process system gains more optimum results than those reported in literature. Computational results indicate that LCA is a very effective algorithm to solve nonlinear programming problems and mixed-integer nonlinear programming problems.A framework is given to solve combinatorial optimization problems and two principles of determining mutation adjacent region are presented. Based on the these two principles, the corresponding mutation adjacent regions are defined for different combinatorial optimization problems, including knapsack problem, traveling salesman problem, pipe line network synthesis, heat exchanger network synthesis and separation sequence synthesis and so on. Solving the above mentioned problems using LCA shows good global search ability, gaining better results for most problems than those using other algorithms. It fully illustrates that LCA has great potential for solving combinatorial optimization problems.