Multi-Objective Optimization in the Digital and AI Eras: A Review of Methodological Evolution and Cross-Domain Applications
- 1School of Economics and Management, Sanming University, No. 25, Ching-Tung Rd., Sanming 365004, China
- 2School of Journalism and Communication, Huaqiao University, No. 266 Chenghua North Road, Jimei District, Xiamen 361021, China
- 3School of Education and Psychology, Minnan Normal University, No. 36, Shì Qian Zhi St., Zhangzhou 363000, China
Abstract
With the rapid advancement of digitalization, data-intensive computing, and artificial intelligence (AI), multi-objective optimization (MOO) has shifted from traditional mathematical modeling and algorithmic solving toward integrated decision-support frameworks that incorporate data-driven learning, decision-maker preferences, and complex contextual factors. However, existing literature reviews still exhibit four major gaps: a lack of integrated cross-period evolutionary perspectives, insufficient links between methodological advances and cross-domain decision values, an incomplete synthesis of emerging challenges in AI-driven MOO, and a scarcity of comprehensive frameworks encompassing resilience, sustainability, equity, and public value. To bridge these gaps, this review synthesizes 140 core analytical publications (2000–2026) indexed in the Web of Science Core Collection to establish a cross-period evolutionary framework, integrate major methodological approaches, evaluate decision values across diverse fields, and outline future research directions in the AI era. We trace the development of MOO across three analytical stages: Information-Based Computing (2000–2010), Data-Intensive Decision-Making (2011–2015), and AI-Driven Optimization (2016–present). Furthermore, we examine the evolution of five core paradigms—mathematical programming, evolutionary multi-objective optimization, multi-criteria decision-making, uncertainty and robust modeling, and AI-assisted and data-driven optimization—across seven application domains. Our findings reveal that the research paradigm of MOO has shifted from computational efficiency and solution-set quality toward data-driven adaptability, dynamic response, trustworthy decision-making, and cross-domain value balance. The primary contribution of this study lies in proposing an integrated “Technology Evolution–Mathematical Methods–Application Value” review framework that aligns methodological features with domain-specific needs, facilitates cross-domain comparisons, and highlights future avenues emphasizing interpretability, reliability, human preferences, and public value.
Related
- A sequential quadratic programming method for optimization with expectation objective functions, deterministic inequality constraints and robust subproblems
- A connectedness result for convex polynomial multi-objective optimization without boundedness
- Representing the Non-dominated Set of Multi-objective Optimization Problems with Polyhedral Ordering Cones
- MPFA: A Pareto Front Approximation Method for Riemannian Bi-objective Optimization
- Simple Extensions of Single-Objective Acquisition Functions and Hedge Strategies for Multi-Objective Bayesian Optimization
BibTeX
@article{lin2026multiobjective,
title = {{Multi-Objective Optimization in the Digital and AI Eras: A Review of Methodological Evolution and Cross-Domain Applications}},
author = {Hung-Lung Lin and Hongxi Shen and Yu-Yu Ma},
journal = {Mathematics},
year = {2026},
month = oct,
doi = {10.3390/math14193602},
url = {https://www.mdpi.com/2227-7390/14/19/3602},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Multi-Objective Optimization in the Digital and AI Eras: A Review of Methodological Evolution and Cross-Domain Applications
AU - Hung-Lung Lin
AU - Hongxi Shen
AU - Yu-Yu Ma
PY - 2026
DA - 2026/10/04
JO - Mathematics
DO - 10.3390/math14193602
UR - https://www.mdpi.com/2227-7390/14/19/3602
AB - With the rapid advancement of digitalization, data-intensive computing, and artificial intelligence (AI), multi-objective optimization (MOO) has shifted from traditional mathematical modeling and algorithmic solving toward integrated decision-support frameworks that incorporate data-driven learning, decision-maker preferences, and complex contextual factors. However, existing literature reviews still exhibit four major gaps: a lack of integrated cross-period evolutionary perspectives, insufficient links between methodological advances and cross-domain decision values, an incomplete synthesis of emerging challenges in AI-driven MOO, and a scarcity of comprehensive frameworks encompassing resilience, sustainability, equity, and public value. To bridge these gaps, this review synthesizes 140 core analytical publications (2000–2026) indexed in the Web of Science Core Collection to establish a cross-period evolutionary framework, integrate major methodological approaches, evaluate decision values across diverse fields, and outline future research directions in the AI era. We trace the development of MOO across three analytical stages: Information-Based Computing (2000–2010), Data-Intensive Decision-Making (2011–2015), and AI-Driven Optimization (2016–present). Furthermore, we examine the evolution of five core paradigms—mathematical programming, evolutionary multi-objective optimization, multi-criteria decision-making, uncertainty and robust modeling, and AI-assisted and data-driven optimization—across seven application domains. Our findings reveal that the research paradigm of MOO has shifted from computational efficiency and solution-set quality toward data-driven adaptability, dynamic response, trustworthy decision-making, and cross-domain value balance. The primary contribution of this study lies in proposing an integrated “Technology Evolution–Mathematical Methods–Application Value” review framework that aligns methodological features with domain-specific needs, facilitates cross-domain comparisons, and highlights future avenues emphasizing interpretability, reliability, human preferences, and public value.
SN - 2227-7390
ER -