Solving the QAP by Two-Stage Graph Pointer Networks and Reinforcement Learning
This work addresses the QAP, a practical optimization problem, with a novel deep learning approach that offers faster solutions than heuristics, though it is incremental in adapting existing methods.
The paper tackles the Quadratic Assignment Problem (QAP), an NP-hard combinatorial optimization challenge, by proposing a two-stage graph pointer network (GPN) with reinforcement learning, achieving semi-optimal solutions on benchmark instances from TSPlib and QAPLIB.
Quadratic Assignment Problem (QAP) is a practical combinatorial optimization problems that has been studied for several years. Since it is NP-hard, solving large problem instances of QAP is challenging. Although heuristics can find semi-optimal solutions, the execution time significantly increases as the problem size increases. Recently, solving combinatorial optimization problems by deep learning has been attracting attention as a faster solver than heuristics. Even with deep learning, however, solving large QAP is still challenging. In this paper, we propose the deep reinforcement learning model called the two-stage graph pointer network (GPN) for solving QAP. Two-stage GPN relies on GPN, which has been proposed for Euclidean Traveling Salesman Problem (TSP). First, we extend GPN for general TSP, and then we add new algorithms to that model for solving QAP. Our experimental results show that our two-stage GPN provides semi-optimal solutions for benchmark problem instances from TSPlib and QAPLIB.