Multipopulation Ant Colony System With Knowledge-Based Local Searches for Multiobjective Supply Chain Configuration

Supply chain management (SCM) is a significant and complex system in a smart city that requires advanced artificial intelligence (AI) and optimization techniques. The multiobjective supply chain configuration (MOSCC) in SCM is to set the optimal configurations for supply chain members to minimize bo...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on evolutionary computation 2022-06, Vol.26 (3), p.512-526
Main Authors: Zhang, Xin, Zhan, Zhi-Hui, Fang, Wei, Qian, Pengjiang, Zhang, Jun
Format: Article
Language:English
Subjects:
Algorithms
Ant colony system (ACS)
Artificial intelligence
Complex systems
Configuration management
Knowledge based systems
knowledge-based local searches (KLSs)
Lead time
multiobjective optimization
Multiple objective analysis
multiple populations for multiple objectives (MPMOs)
Optimization
Optimization techniques
Search problems
Searching
Smart cities
smart city
Sociology
Statistics
supply chain configuration (SCC)
Supply chain management
Supply chains
Citations: Items that this one cites
Items that cite this one
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Supply chain management (SCM) is a significant and complex system in a smart city that requires advanced artificial intelligence (AI) and optimization techniques. The multiobjective supply chain configuration (MOSCC) in SCM is to set the optimal configurations for supply chain members to minimize both the cost of goods sold ( CoGS ) and the lead time ( LT ). Although some algorithms have been proposed for the MOSCC, they do not make the best use of the problem-related knowledge and cannot perform well on the large-scale instances with many members and configuration options. Therefore, this article proposes a multipopulation ant colony system with knowledge-based local searches (MPACS-KLSs). First, the multiobjective algorithm is based on the multiple populations for multiple objectives framework. Two ant colonies are used to separately minimize CoGS and LT , which helps to search in the biobjective space sufficiently. Second, with the considerations of the problem-related knowledge, a priority-based solution construction method, a rank-based heuristic strategy, and an objective-oriented global pheromone updating strategy are proposed. Third, to speed up the convergence, especially for large-scale MOSCC instances, two knowledge-based local searches are designed to minimize CoGS and LT of solutions, respectively. Exhaustive experiments are conducted on both the instances from the real life and the randomly generated instances with different problem scales. The results show that MPACS-KLS is superior to the contestant algorithms, especially on the large-scale MOSCC instances, which significantly extends the AI and optimization techniques in practical applications of the smart city.
ISSN:1089-778X
1941-0026
DOI:10.1109/TEVC.2021.3097339