Title :
A multivalued Boltzmann machine
Author :
Lin, C.T. ; Lee, C.S.G.
Author_Institution :
Sch. of Electr. Eng., Purdue Univ., West Lafayette, IN, USA
Abstract :
The original Hopfield network and the Boltzmann machine are generalized based on the XY spin glass model and the discrete planar spin glass model to develop a multivalued Boltzmann machine. A general Hamiltonian is defined to realize this multivalued machine, and three distributed computational architectures with evolving rules are proposed to minimize this Hamiltonian. The machine was applied to the mobile robot navigation problem using an artificial magnetic field approach. Computer simulations were conducted and they verified the validity of the proposed system
Keywords :
neural nets; spin Hamiltonians; Hopfield neural network; XY spin glass model; artificial magnetic field; discrete planar spin glass model; general Hamiltonian minimization; mobile robot navigation; multivalued Boltzmann machine; Analog computers; Associative memory; Computational modeling; Glass; Mobile robots; Motion planning; Neural networks; Neurons; Simulated annealing; Stochastic processes;
Conference_Titel :
Neural Networks, 1991. 1991 IEEE International Joint Conference on
Print_ISBN :
0-7803-0227-3
DOI :
10.1109/IJCNN.1991.170772
