Dynamics of Alopex Process: Application to Optimization Problems

The Alopex process has been applied to a variety of problems in which maxima or minima of functions of many parameters are to be determined. We present here different forms of the algorithm and a simple theory of the dynamics of the process. Computer simulations of several applications are presented to illustrate the versatility of the method. An artificial visual system is described in which partial stimuli are completed and patterns are created out of noise with only a scalar cost function to guide the process.

[1]  N. Metropolis,et al.  Equation of State Calculations by Fast Computing Machines , 1953, Resonance.

[2]  Geoffrey E. Hinton,et al.  OPTIMAL PERCEPTUAL INFERENCE , 1983 .

[3]  E Harth,et al.  The inversion of sensory processing by feedback pathways: a model of visual cognitive functions. , 1987, Science.

[4]  E. Harth,et al.  Brainstem control of sensory information: a mechanism for perception. , 1985, International journal of psychophysiology : official journal of the International Organization of Psychophysiology.

[5]  J. Hopfield,et al.  Computing with neural circuits: a model. , 1986, Science.

[6]  E Harth,et al.  Alopex: a stochastic method for determining visual receptive fields. , 1974, Vision research.

[7]  C. D. Gelatt,et al.  Optimization by Simulated Annealing , 1983, Science.