RM-MOCO: A Fast-Solving Model for Neural Multi-Objective Combinatorial Optimization Based on Retention

Multiobjective combinatorial optimization (MOCO) problems have a wide range of applications in the real world. Recently, learning-based methods have achieved good results in solving MOCO problems. However, most of these methods use attention mechanisms and their variants, which have room for further...

Full description

Saved in:
Bibliographic Details
Main Authors: Huiqing Wei, Fei Han, Qing Liu, Henry Han
Format: Article
Language:English
Published: Tsinghua University Press 2025-06-01
Series:Complex System Modeling and Simulation
Subjects:
Online Access:https://www.sciopen.com/article/10.23919/CSMS.2024.0029
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849762617634062336
author Huiqing Wei
Fei Han
Qing Liu
Henry Han
author_facet Huiqing Wei
Fei Han
Qing Liu
Henry Han
author_sort Huiqing Wei
collection DOAJ
description Multiobjective combinatorial optimization (MOCO) problems have a wide range of applications in the real world. Recently, learning-based methods have achieved good results in solving MOCO problems. However, most of these methods use attention mechanisms and their variants, which have room for further improvement in the speed of solving MOCO problems. In this paper, following the idea of decomposition strategy and neural combinatorial optimization, a novel fast-solving model for MOCO based on retention is proposed. A brand new calculation of retention is proposed, causal masking and exponential decay are deprecated in retention, so that our model could better solve MOCO problems. During model training, a parallel computation of retention is applied, allowing for fast parallel training. When using the model to solve MOCO problems, a recurrent computation of retention is applied, enabling quicker problem-solving. In order to make our model more practical and flexible, a preference-based retention decoder is proposed, which allows generating approximate Pareto solutions for any trade-off preferences directly. An industry-standard deep reinforcement learning algorithm is used to train RM-MOCO. Experimental results show that, while ensuring the quality of problem solving,the proposed method significantly outperforms some other methods in terms of the speed of solving MOCO problems.
format Article
id doaj-art-0f15ecc1e90e4eba84724e45906b2768
institution DOAJ
issn 2096-9929
2097-3705
language English
publishDate 2025-06-01
publisher Tsinghua University Press
record_format Article
series Complex System Modeling and Simulation
spelling doaj-art-0f15ecc1e90e4eba84724e45906b27682025-08-20T03:05:42ZengTsinghua University PressComplex System Modeling and Simulation2096-99292097-37052025-06-015212513710.23919/CSMS.2024.0029RM-MOCO: A Fast-Solving Model for Neural Multi-Objective Combinatorial Optimization Based on RetentionHuiqing Wei0Fei Han1Qing Liu2Henry Han3School of Computer Science and Communication Engineering, Jiangsu University, Zhenjiang 212013, ChinaSchool of Computer Science and Communication Engineering, Jiangsu University, Zhenjiang 212013, ChinaSchool of Electronic and Information Engineering, West Anhui University, Lu’an 237005, ChinaSchool of Engineering and Computer Science, Baylor University, Waco, TX 76798, USAMultiobjective combinatorial optimization (MOCO) problems have a wide range of applications in the real world. Recently, learning-based methods have achieved good results in solving MOCO problems. However, most of these methods use attention mechanisms and their variants, which have room for further improvement in the speed of solving MOCO problems. In this paper, following the idea of decomposition strategy and neural combinatorial optimization, a novel fast-solving model for MOCO based on retention is proposed. A brand new calculation of retention is proposed, causal masking and exponential decay are deprecated in retention, so that our model could better solve MOCO problems. During model training, a parallel computation of retention is applied, allowing for fast parallel training. When using the model to solve MOCO problems, a recurrent computation of retention is applied, enabling quicker problem-solving. In order to make our model more practical and flexible, a preference-based retention decoder is proposed, which allows generating approximate Pareto solutions for any trade-off preferences directly. An industry-standard deep reinforcement learning algorithm is used to train RM-MOCO. Experimental results show that, while ensuring the quality of problem solving,the proposed method significantly outperforms some other methods in terms of the speed of solving MOCO problems.https://www.sciopen.com/article/10.23919/CSMS.2024.0029multiobjective combinatorial optimizationlearning-based methodretention modeldeep reinforcement learning
spellingShingle Huiqing Wei
Fei Han
Qing Liu
Henry Han
RM-MOCO: A Fast-Solving Model for Neural Multi-Objective Combinatorial Optimization Based on Retention
Complex System Modeling and Simulation
multiobjective combinatorial optimization
learning-based method
retention model
deep reinforcement learning
title RM-MOCO: A Fast-Solving Model for Neural Multi-Objective Combinatorial Optimization Based on Retention
title_full RM-MOCO: A Fast-Solving Model for Neural Multi-Objective Combinatorial Optimization Based on Retention
title_fullStr RM-MOCO: A Fast-Solving Model for Neural Multi-Objective Combinatorial Optimization Based on Retention
title_full_unstemmed RM-MOCO: A Fast-Solving Model for Neural Multi-Objective Combinatorial Optimization Based on Retention
title_short RM-MOCO: A Fast-Solving Model for Neural Multi-Objective Combinatorial Optimization Based on Retention
title_sort rm moco a fast solving model for neural multi objective combinatorial optimization based on retention
topic multiobjective combinatorial optimization
learning-based method
retention model
deep reinforcement learning
url https://www.sciopen.com/article/10.23919/CSMS.2024.0029
work_keys_str_mv AT huiqingwei rmmocoafastsolvingmodelforneuralmultiobjectivecombinatorialoptimizationbasedonretention
AT feihan rmmocoafastsolvingmodelforneuralmultiobjectivecombinatorialoptimizationbasedonretention
AT qingliu rmmocoafastsolvingmodelforneuralmultiobjectivecombinatorialoptimizationbasedonretention
AT henryhan rmmocoafastsolvingmodelforneuralmultiobjectivecombinatorialoptimizationbasedonretention