Logistics scheduling planning method based on graph neural network and reinforcement learning

Abstract

The invention discloses a logistics scheduling planning method based on a graph neural network and reinforcement learning. The method comprises the following steps: step 1, constructing a complete solution of a vehicle path planning problem instance; step 2, selecting a disturbance controller or a lifting controller by the element controller; after the lifting controller is selected, the lifting operator set forms an action space of the lifting controller; training the graph neural network in the action space; step 3, carrying out de-lifting; step 4, if the meta controller selects a disturbance controller, the disturbance controller randomly selects a disturbance operator to disrupt and reconstruct a feasible solution, and then iterative lifting is carried out to find an optimal solution; and step 5, selecting a solution with the minimum total path length in all visited feasible solutions in the lifting and disturbance process as an optimal solution and a final output solution of the whole algorithm. Compared with the prior art, the optimal solution of the given problem can be efficiently searched, and the method has practical significance for planning problems such as logistics and order distribution.

Description

technical field [0001] The invention relates to a graph neural network and reinforcement learning technology, in particular to a method for controlling the selection of a heuristic operator by combining a graph neural network and a strategy gradient algorithm in reinforcement learning. Background technique [0002] NP-hard combinatorial optimization problems are a class of integer-constrained optimization problems that are difficult to solve in large-scale optimization. Robust approximation algorithms for NP-hard combinatorial optimization problems have many practical applications, such as transportation, supply chain, energy, finance and The backbone of modern industries such as scheduling. A typical example is the Traveling Salesman Problem algorithm (Traveling SalesmanProblem, TSP), in this algorithm, given a graph, the goal is to search the permutation space, and find the minimum total edge weight and (tour length) optimal node sequence. TSP and its variants have count...

AI Technical Summary

Problems solved by technology

[0003] While most successful machine learning techniques fall into the domain of supervised learning, i.e. learning a mapping from training inputs to outputs, supervised learning is not suitable for most combinatorial optimization problems because one cannot obtain optimal labels

At the same time, traditional reinforcement learning methods are often limited to training and solving problems with fixed node sizes due to the fixed size of network parameters when solving path planning problems.

Images