首页>
外国专利>
PARTS ARRANGEMENT OPTIMIZING ALGORITHM IN NEURAL NETWORK
PARTS ARRANGEMENT OPTIMIZING ALGORITHM IN NEURAL NETWORK
展开▼
机译:神经网络中的零件布置优化算法
展开▼
页面导航
摘要
著录项
相似文献
摘要
PURPOSE:To obtain a parts arrangement optimizing algorithm which can be applied to a large scale problem with a large number of parts by updating weight by means of learning through the use of N-number of neurons and 2N- number of weights to obtain optimum arrangement. CONSTITUTION:N-number of neurons are used for the problem of N-number of the parts, and the respective neurons are allocated to the respective parts one by one. At the time of two-dimensionally arranging the parts, two input lines 2a and 2b re allocated to the neurons 1a-1c and a position coordinate is inputted to the respective input lines 2a and 2b as input. The neurons have two-dimensional weights 3a-3f in correspondence with the respective input lines. The position coordinate where the parts are arranged is inputted to the neuron, and the optimum and adaptive neuron is selected under any reference to the input so as to update the weight of the neuron. The parts allocated to the selected optimum and adaptive neuron are arranged in a parts arrangement position and one parts arrangement is obtained by repeating the above cycle.
展开▼