基于BP算法的IFPN参数优化方法

针对直觉模糊Petri 网(IFPN) 模型自学习能力差的缺陷, 将神经网络中的BP 误差反传算法引入IFPN 模型的参数寻优过程, 提出一种基于此的参数优化方法. 该算法通过建立变迁点燃和直觉模糊推理的近似连续函数, 摆脱了参数对经验的依赖, 更加符合实际系统的需求, 同时使得IFPN 具有较强的泛化能力和自适应功能, 推理结果更加准确可信. 最后通过典型实例验证了该参数优化方法的有效性和优越性.

Method for parameters optimization of IFPN based on BP algorithm

In order to improve the self-learning capability of intuitionistic fuzzy Petri nets(IFPN), a novel parametersoptimization method is proposed, in which the back propagation algorithm of neural net is introduced to the parameters-optimized process of IFPN. By constructing the approximate continuous function of transition firing and intuitionistic fuzzyreasoning, the method makes the parameters get rid of the dependence upon experience, which makes the parameters adjustthe fact instance better. Meanwhile, the IFPN model can own better generalization performance and self-adjusting ability,and the reasoning results are more accurate and reliable as well. Finally, the classical instance verifies the effectiveness andsuperiority of the proposed method.