Search

High-order neural networks can be considered as an expansion of Hopfield neural networks, and have stronger approximation property and faster convergence rate. However, in practice high order network is seldom to be used to solve combinatorial optimization problem. In this paper crossbar switch problem, which is an NP-complete problem, is used as an example to demonstrate how to use high order discrete Hopfield neural network to solve engineering optimization problems. The construction method of energy function and the neural computing algorithm are presented. It is also discussed the method how to speed the convergence and escape from local minima. Experimental results show that high order network has a quick convergence speed, and outperforms the traditional discrete Hopfield network. [ABSTRACT FROM AUTHOR]

Dynamic programming has provided a powerful approach to solve optimization problems, but its applicability has sometimes been limited because of the high computational effort required by the conventional algorithms. This paper presents an association between Hopfield networks and genetic algorithms, which cover extensive search spaces and guarantee the convergence of the system to the equilibrium points that represent feasible solutions for dynamic programming problems. [ABSTRACT FROM AUTHOR]

In this paper, we explore the recognition of polyphone. The cognition process is complex, which needs other additional information, otherwise it may cause uncertainty in decision. Recent research is almost focused on phonetics, while we plan to explore the question with neural networks. H. Haken used synergetic neural network to discuss the recognition of ambivalent patterns and the evolution equation of order parameters can interpret the oscillation in perception. Based on his idea, we argue that the process of cognition is phase transformation. Then we apply Hopfield network (associative memory network) with depressing synapse to simulate the recognition process. With our model, a Chinese polyphone is demonstrated. The result supports our interpretation strongly. [ABSTRACT FROM AUTHOR]

Formation of agents is of recent interest in computer sciences, robotics and control systems. Several goal formation strategies may be of interest according to the sensory capabilities of the agents. This paper addresses the formation of mobile agents in absolute positioning without order. In this paper a control system based on Hopfiled neural networks is proposed. The paper summarizes the control system and describes a JAVA-based application developed to visualize the control system behavior. This interactive animation tool improves the understanding of and intuition for a number of aspects dealing with the formation of agents such as agents dynamics. This tool allows to directly manipulate graphical representation of the systems such as initial configuration of the agents, and get instant feedback on the effects. [ABSTRACT FROM AUTHOR]

The linear model of flexible two-wheeled self-balancing robot that has been proved is used in the paper. The lumbar of robot is the spring. The discrete Hopfield neural network is used for controlling robot΄s posture based on association study. Study of biological control method to realize flexible robot posture control in the adaptive, self-organizing capacity, the reasonable energy function is defined for the non-linear and strong coupling robot, then design the Hopfield controller, used the dynamic function of flexible robot, which satisfied the need of robot΄s dynamic process. Analyze the convergence of the controller. Simulation experiments show the effectiveness of the method and proved the stability of system. Validity and rationality of the system controller designed are verified through the performance experiments of the prototype and analyzed the experiment result in detail. The robot like the human being not only structure, but also physiological intelligence. [ABSTRACT FROM AUTHOR]

This paper presents a neurogenetic approach for solving nonlinear programming problems. Genetic algorithm must its popularity to make possible cover nonlinear and extensive search spaces. Neural networks with feedback connections provide a computing model capable of solving a large class of optimization problems. The association of a modified Hopfield network with genetic algorithm guarantees the convergence of the system to the equilibrium points, which represent feasible solutions for nonlinear programming problems. [ABSTRACT FROM AUTHOR]

This paper constructs a balanced training used in the hairy network [1] to balance the vulnerable memory parts and improve the memory. It provides a perspective view on the geometrical structure of memory patterns. This training fixes many drawbacks of the Hopfield network, such as loading capacity, limit cycle and error tolerance. [ABSTRACT FROM AUTHOR]