论文信息 - Efficient energy landscape transformation in the problem of binary minimization

Efficient energy landscape transformation in the problem of binary minimization

A problem of quadratic functional minimization in a discrete space is considered. It is shown that the transformation of a functional by modification of its matrix can significantly accelerate a procedure of a random search. As example we chose two well-known local optimization algorithms: Hopfield neural-network dynamics and Kernighan-Lin algorithm. The proposed method of functional transformation improves efficiency of the both algorithms by many times.

Iakov M. Karandashev | Boris Kryzhanovsky

[1] B. Kryzhanovsky. Expansion of a matrix in terms of external products of configuration vectors , 2008, Optical Memory and Neural Networks.

[2] H. Rieger,et al. New Optimization Algorithms in Physics , 2004 .

[3] Jiye Liang,et al. A deterministic annealing algorithm for approximating a solution of the min-bisection problem , 2009, Neural Networks.

[4] Brian W. Kernighan,et al. An efficient heuristic procedure for partitioning graphs , 1970, Bell Syst. Tech. J..

[6] H. Rieger,et al. New Optimization Algorithms in Physics: HARTMANN:NEW OPT.ALG.PHY. O-BK , 2004 .

[7] J J Hopfield,et al. Neural networks and physical systems with emergent collective computational abilities. , 1982, Proceedings of the National Academy of Sciences of the United States of America.

[8] Sompolinsky,et al. Spin-glass models of neural networks. , 1985, Physical review. A, General physics.

[9] O. Martin,et al. Hierarchical approach for computing spin glass ground states. , 2001, Physical review. E, Statistical, nonlinear, and soft matter physics.

[10] B. V. Kryzhanovskii,et al. A relation between the depth of a local minimum and the probability of its detection in the generalized hopfield model , 2005 .