A Self-Adaptive Hybrid Optimization Algorithm for Solving Consecutive Reaction Problem

Abstract

The actual temperature control for consecutive reaction problem is a complex optimization problem. Genetic algorithms (GA) is a metaheuristic inspired by imitating the processes observed during natural evolution. It has a strong global search ability and less computation time, but it exists the premature convergence and poor stability. Ant colony optimization (ACO) is a metaheuristic inspired by imitating the behavior of real ants. It has the robustness and parallel computation, but it exists the slow convergence speed and stagnation phenomenon. In this paper, a new genetic and ant colony self-adaptive hybrid (NGASAH) algorithm based on the chaotic searching strategy, multi-populations and self-adaptive parameter control strategies is presented. In the proposed NGASAH algorithm, the chaotic searching strategy is used to avoid the optimal solution. The strategy of the multiple populations is used to avoid to converge to a local extreme point of all particles. The strategy of self-adaptive parameter control is used to dynamically balance the local search ability and the global ability, and improve the convergence speed. The actual temperature control of consecutive reaction problem is used to test the validity of the NGASAH algorithm. The experiment results show that the NGASAH algorithm can obtain the global search ability and the faster convergence speed in solving the complex optimization problems.