Electronic neural network for dynamic resource allocation

A VLSl implementable neural etwork architecture for dynamic assignment is presented. The dynamic assignment or resource allocation problems involve assigning-members of one set (e. g. resources) to those of an other (e. g. consumers) such that the global "cost" of the associations is minimized. The network consists of a matrix of sigmoidal processing elements (neurons), where the rows of the matrix represent resources and columns represent consumers. Unlike previous neural implementations however, association costs are applied directly to the neurons, reducing connectivity of the network to VLSI-compatible O(number of neurons). Each row (and column) has an additional neuron associated with it to independently oversee activations of all the neurons in each row (and each column), providing a programmable "k- w i n ne r - t a ke -a I I" f u nct i o n . This function simultaneously enforces blocking (excitatory / inhibitory) constraints during convergence to control the number of active elements in each row and column within desired boundary conditions. Simulations show that the network when implemented in fully parallel VLSl hardware would offer optimal (or a near-optimal) solution within only a fraction of a millisecond, for problems up to 128 resources and 128 consumers, orders of magnitude faster than conventional computing or heuristic search methods. Furthermore, this network would provide, for the first time, a unique capability to solve allocation problems with arbitrary many-to-many assignment constraints in real time. 1 ntroduct ion

[1]  A. Thakoor,et al.  Design of parallel hardware neural network systems from custom analog VLSI 'building block' chips , 1989, International 1989 Joint Conference on Neural Networks.

[3]  Yaser S. Abu-Mostafa,et al.  On the K-Winners-Take-All Network , 1988, NIPS.

[4]  Stephen Grossberg,et al.  A massively parallel architecture for a self-organizing neural pattern recognition machine , 1988, Comput. Vis. Graph. Image Process..

[5]  A. Moopenn,et al.  A neural network for Euclidean distance minimization , 1988, IEEE 1988 International Conference on Neural Networks.

[6]  J J Hopfield,et al.  Neurons with graded response have collective computational properties like those of two-state neurons. , 1984, Proceedings of the National Academy of Sciences of the United States of America.

[7]  Gene A. Tagliarini,et al.  A Neural-Network Solution to the Concentrator Assignment Problem , 1987, NIPS.

[8]  Timothy X. Brown,et al.  Competitive neural architecture for hardware solution to the assignment problem , 1991, Neural Networks.

[9]  Yasuji Sawada,et al.  Associative memory network composed of neurons with hysteretic property , 1990, Neural Networks.

[10]  J. Munkres.  ALGORITHMS FOR THE ASSIGNMENT AND TRANSIORTATION tROBLEMS* , 1957 .

[11]  Alexander Moopenn,et al.  Digital-Analog Hybrid Synapse Chips for Electronic Neural Networks , 1989, NIPS.

[12]  Jin Luo,et al.  CMOS UV-writable non-volatile analog storage , 1991 .

[13]  C. Portmann,et al.  Practical design and analysis of a simple 'neural' optimization circuit , 1989 .