Approximating metaheuristics with homotopic recurrent neural networks
This topic contains 0 replies, has 1 voice, and was last updated by arXiv 1 year, 3 months ago.

Approximating metaheuristics with homotopic recurrent neural networks
Much combinatorial optimisation problems constitute a nonpolynomial (NP) hard optimisation problem, i.e., they can not be solved in polynomial time. One such problem is finding the shortest route between two nodes on a graph. Metaheuristic algorithms such as $A^{*}$ along with mixedinteger programming (MIP) methods are often employed for these problems. Our work demonstrates that it is possible to approximate solutions generated by a metaheuristic algorithm using a deep recurrent neural network. We compare different methodologies based on reinforcement learning (RL) and recurrent neural networks (RNN) to gauge their respective quality of approximation. We show the viability of recurrent neural network solutions on a graph that has over 300 nodes and argue that a sequencetosequence network rather than other recurrent networks has improved approximation quality. Additionally, we argue that homotopy continuation — that increases chances of hitting an extremum — further improves the estimate generated by a vanilla RNN.
Approximating metaheuristics with homotopic recurrent neural networks
by Alessandro Bay, Biswa Sengupta
https://arxiv.org/pdf/1709.02194v1.pdf
You must be logged in to reply to this topic.