Application of Self-adaptive Master-slave Parallel Genetic Algorithm to Interval Nonlinear Programming

This paper considers the nonlinear programming problem of interval parameters.A general interpretation formulation of nonlinear programming under uncertainty is proposed with the introduction of decision making risk factors.To solve this formulation,this paper presents a self-adaptive master-slave parallel genetic algorithm,which meets the real-time requirements of large scale optimization problem and has the capability of global convergence.Compared with the traditional master-slave parallel genetic algorithms,the presented algorithm can efficiently solve the problem of unbalanced distribution of computational load among the slave computers by dynamically adjusting computational load of the slave computers.Simulation result proves the feasibility of the presented algorithm.