Abstract
A novel computational paradigm with high-dimensional analog chaotic neuro-dynamics is explored.
A hardware prototype, which combines analog computations with chaotic neuro-dynamics and digital computation through algorithms, is constructed using analog chaotic-neuron integrated-circuits.
In the system, 300-dimensional analog chaotic neuro-dynamics drive a tabu-search algorithm. We demonstrate experimentally that the prototype system efficiently solves
quadratic assignment problems (QAPs) as benchmarks through physical chaotic dynamics.
We also qualitatively analyze the underlying mechanism of the highly parallel and collective analog computations by observing global and local dynamics.
Furthermore, we introduce temporal mutual information to quantitatively evaluate the system dynamics.
In addition, we propose a synchronous update algorithm with modified neuron model suitable for a massively parallel hardware system in order to effectively solve large-size QAPs.
The experimental and simulation results confirm the validity and efficiency of the proposed computational paradigm with the physical analog chaotic neuro-dynamics.
We may assume that this hybrid system would be analogous to the information processing in the brain based on the interaction between conscious and subconscious processes.
That is, the heuristic algorithm and the analog chaotic neuro-dynamics correspond to the conscious and subconscious processes, respectively.
