Research On Mechanism Kinematic Chain Isomorphism Identification Based On Genetic Hopfield Neural Network

Mechanism kinematic chain isomorphism identification is a key issue that must be solved in the synthesis process of kinematic chain structure,which is of great significance to the mechanism innovation design.Because this problem has not been solved effectively,searching for an accurate and efficient a method for kinematic chain isomorphism identification has been one of the goals in the development of mechanism science.In this paper,graph theory is introduced into the study of mechanism,and the topological graph is used to represent the mechanism kinematic chain,which has a one-to-one correspondence with the mechanism kinematic chain.In order to facilitate computer processing,according to the mathematical description of the topological graph,all the structural information of the topological graph is represented by an adjacency matrix.Based on the concept of graph isomorphism and description model of mechanism kinematic chain,the criterion of isomorphism detection for mechanism kinematic chain is proposed,which lays a theoretical foundation for the problem of mechanism kinematic chain isomorphism identification using optimization algorithms.On this basis,the Hopfield neural network algorithm is applied to isomorphism problem in kinematic chains,so as to realize the automatic identification of the kinematic chain isomorphism.The problem of kinematic chain isomorphism identification is mapped to the Hopfield neural network.According to the characteristics of isomorphism problem in kinematic chains,a Hopfield neural network model for kinematic chain isomorphism identification is established.Finally,the above Hopfield neural network algorithm is used to simulate and test four different types of kinematic chains.The test results show that this algorithm is feasible and efficient for solving isomorphism problem in kinematic chains.Aiming at some deficiencies in Hopfield neural network algorithm for solving the problem of mechanism kinematic chain isomorphism identification,a genetic algorithm is introduced,Hopfield neural network algorithm and genetic algorithm arecombined effectively,and an isomorphism identification method for mechanism kinematic chain based on genetic Hopfield network is proposed.Hopfield neural network algorithm and genetic algorithm are complementary to a certain extent,so as to improve the performance of the entire algorithm.Some parts of genetic algorithm in the genetic Hopfield neural network model are improved.The selection operator of deterministic sampling selection plus optimal reservation selection strategy is used to enhance the search efficiency of genetic algorithm.The local search operator is added so that the genetic algorithm has a certain fine-tuning ability.In the genetic Hopfield neural network model with kinematic chain isomorphism identification,the initial state of the Hopfield neural network is firstly optimized by using genetic algorithm,which makes the initial state of the Hopfield neural network achieve a better state quickly,and then it uses the Hopfield neural network algorithm to iterate initial state after optimization until network state is stable.Simulation experiments show that the hybrid algorithm has a better solution than the Hopfield neural network algorithm.