论文部分内容阅读
作为一类重要的模糊规划问题,模糊期望值规划(Fuzzy Expect Value Model,FEVM)在理论和工程应用方面有着广泛的应用背景,为了探求更为高效的FEVM求解方法,模糊模拟在本文中被用来产生神经网络的训练样本,然后训练BP神经网络以逼近模糊函数,再应用微粒群算法并以逼近模糊函数的神经网络作为适应值计算和实现检验可行解,从而提出了一种混合智能算法来求解FEVM问题。实验对照后,说明了算法的有效性和高的计算效率,为FEVM问题的求解提供了一种有效的途径。
As an important fuzzy programming problem, Fuzzy Expect Value Model (FEVM) has a wide range of applications in theory and engineering applications. In order to explore a more efficient FEVM solution, fuzzy simulation is used in this paper A training sample of neural network is generated, BP neural network is trained to approximate fuzzy function, and then a particle swarm optimization algorithm is used to compute and verify the feasible solution by using the neural network approximation fuzzy function as a fitness value. A hybrid intelligent algorithm is proposed to solve this problem FEVM problem. The experimental comparison shows the effectiveness of the algorithm and the high computational efficiency, which provides an effective way to solve the FEVM problem.