A workshop production scheduling method and system based on deep reinforcement learning and multimodal multi-objective computation

By combining deep reinforcement learning and multimodal multi-objective evolutionary algorithms with serial-parallel mechanisms and neighborhood dominance relationships, the shortcomings of multimodal multi-objective optimization algorithms in maintaining diversity and processing decision space are addressed, achieving more efficient multi-objective optimization results.

CN121094419BActive Publication Date: 2026-05-26GANNAN NORMAL UNIV
0 Cites 0 Cited by

Patent Information

Application Number
CN202511198069.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-08-26
Publication Date
2026-05-26
Estimated Expiration
2045-08-26

AI Technical Summary

Technical Problem

Existing multimodal multi-objective optimization algorithms have shortcomings in maintaining diversity, handling decision space, and dynamic optimization. These shortcomings lead to the algorithms being overly sensitive to problem characteristics, failing to accurately identify the diverse needs of the population and the global optimization direction, and having insufficient evaluation accuracy in dynamic environments.

Method used

A multimodal, multi-objective evolutionary algorithm guided by deep reinforcement learning is adopted. The Q-value of actions is evaluated through a deep Q-network. The algorithm optimizes the diversity and convergence of the population by combining a serial-parallel mechanism and an adaptive search space adjustment strategy. The population diversity is evaluated using a neighborhood-dominant update method and crowding distance, thereby enhancing the performance of the algorithm in complex multimodal, multi-objective optimization problems.

Benefits of technology

It significantly improves the overall performance of the algorithm in multimodal and multi-objective optimization problems. By using the adaptive selection operator and neighborhood dominance ranking mechanism of the deep Q-network, it improves the diversity and convergence of the population, enhances the ability to handle high-dimensional decision variables, and reduces the sensitivity of the algorithm.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

This invention discloses a workshop production scheduling method and system based on deep reinforcement learning and multimodal multi-objective computation. The method includes the following steps: obtaining a population size N, an objective function, and a maximum number of iterations; randomly generating an initial population based on the population size and the objective function, defining the state of the initial population, and constructing a deep Q-network; setting a counter and comparing the counter with a stagnation check period; adaptively selecting operators for the initial population using the deep Q-network to generate a child population; generating three temporary files based on the child population and merging them to obtain a merged file; performing environment selection on the merged file to obtain a new population and calculating state-reward pairs; updating the deep Q-network every 50 iterations based on the state-reward pairs; re-comparing the counter with the stagnation check period, and outputting the scheduling result if the counter reaches the stagnation check period.
Need to check novelty before this filing date? Find Prior Art