一种适于多目标稳态优化决策的伪权向量方法

New Method Based on Pseudo-weighted Vector for Decision-making of Multi-objective Stable State Optimization

伪权向量法是解决多目标决策问题的一种常见方法,但由于其决策函数中缺少对决策变量调整幅值的约束,易造成因决策变量(通常为控制回路的设定值)变动幅度过大而影响被控过程稳定性的后果,因而限制了该方法在过程控制系统稳态优化中的应用。对此,提出了一种适合于过程控制系统稳态优化的伪权向量多目标决策方法。该方法综合考虑了决策者偏好以及决策变量的变化情况,为控制系统的稳态优化提供了一种适合实际的决策方法。仿真结果验证了该方法的有效性和实用性。

Pseudo-weighted vector is a normal method to solve the decision-making question of multi-objective optimization. But the decision function doesn＇t include the constrained of decision variable. The wide changing magnitude of decision variable （generally, the initialization of control variable） maybe affect the stability of the control system. So this deficiency limits its application. In order to avoid this defect, an improved decision-making method based on pseudo-weighted vector was proposed for the process control system stable state optimization. The decision preferential information and the change of decision variablewere all considered to fit the stable state optimization system. The simulation results show that the improved method is simple, efficient and practicable.