Method and system for solving Hamiltonian shortest path based on neural combinatorial optimizer

• Main Authors: JIANG ZENGHUI, GUO HONGWEI, LIANG HELAN, LI FANCHANG
• Format: Patent
• Language: Chinese; English
• Published: 23.01.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FORADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORYOR FORECASTING PURPOSES; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE,COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTINGPURPOSES, NOT OTHERWISE PROVIDED FOR

The invention provides a method and system for solving the shortest Hamiltonian path based on a neural combinatorial optimizer, and relates to the technical field of intelligent calculation and path planning, and the method comprises the steps: designing a shortest Hamiltonian path problem framework; designing a neural combination optimizer model based on the encoder-decoder structure; hamiltonian problem instances used for training the neural combination optimizer are randomly generated, and data of all the problem instances are preprocessed through a minimum-maximum normalization and rotation coordinate transformation strategy; inputting the processed data into the neural combination optimizer model for training, and performing optimization by adopting a bidirectional strategy gradient function to obtain a trained neural combination optimizer model; and solving the shortest Hamiltonian path problem containing N nodes by using the trained neural combination optimizer model to obtain an optimal Hamiltonian pa