Model Predictive Control for Neuromimetic Quantized Systems

TL;DR

This paper introduces a model predictive control approach for neuromimetic quantized systems, ensuring stability and optimal emulation through a combination of MPC, neural network training, and sphere decoding algorithms.

Contribution

It presents a novel MPC-based method for optimal quantization in neuromimetic systems, incorporating neural network training and sphere decoding for improved solution accuracy.

Findings

Guarantees asymptotic stability during emulation

Effective optimization of large discrete input systems

Integration of neural networks with sphere decoding enhances solution quality

Abstract

Based on our recent research on neural heuristic quantization systems, we propose an emulation problem consistent with the neuromimetic paradigm. This optimal quantization problem can be solved with model predictive control (MPC) by deriving the conditions under which the quantized system can guarantee (asymptotic) stability during emulation by optimizing a Lyapunov-like objective function. The neuromimetic model features large numbers of discrete inputs, and the optimization involves integer variables. The approach in the paper begins by solving an optimization using model predictive control (MPC) and then using a neural network to train the data generated in this process and applying Fincke and Pohst's sphere decoding algorithm to narrow down the search for the optimal solution.