论文部分内容阅读
针对一般人工神经网络不能用于求解包含矛盾的约束满足问题（ＣＳＰ）的不足，本文依据神经网络的逻辑分析理论，提出了一个约束满足神经网络（ＣＳＮＮ）．ＣＳＮＮ体现了生物神经系统中的突触的控制原理，它由一个基本神经网和一个控制系统组成；基本网的作用是提供必要的约束，控制系统的作用不但能使约束成为自适应的，而且依据具体问题能够动态地、自动地删除基本解中的矛盾．
Aiming at the deficiency that general artificial neural network can not be used to solve contradictory Constraint Satisfaction Problem (CSP), this paper proposes a Constraint Satisfaction Neural Network (CSNN) based on the theory of neural network. CSNN embodies the principle of synaptic control in the biological nervous system. It consists of a basic neural network and a control system. The function of the basic network is to provide necessary constraints. The function of the control system not only makes the constraints self-adaptive, but also Based on specific problems, the contradictions in the basic solution can be deleted dynamically and automatically.