A Review: Machine Learning for Combinatorial Optimization Problems in Energy Areas

Abstract

Combinatorial optimization problems (COPs) are a class of NP-hard problems with great practical significance. Traditional approaches for COPs suffer from high computational time and reliance on expert knowledge, and machine learning (ML) methods, as powerful tools have been used to overcome these problems. In this review, the COPs in energy areas with a series of modern ML approaches, i.e., the interdisciplinary areas of COPs, ML and energy areas, are mainly investigated. Recent works on solving COPs using ML are sorted out firstly by methods which include supervised learning (SL), deep learning (DL), reinforcement learning (RL) and recently proposed game theoretic methods, and then problems where the timeline of the improvements for some fundamental COPs is the layout. Practical applications of ML methods in the energy areas, including the petroleum supply chain, steel-making, electric power system and wind power, are summarized for the first time, and challenges in this field are analyzed.

Machine Learning's Influence on Supply Chain and Logistics Optimization in the Oil and Gas Sector: A Comprehensive Analysis

“…Oil and gas companies have complex supply chains with data spread across different systems and formats, making data integration challenging (Tirkolaee, et. al., 2021, Yang, et. al., 2022.…”

An Integrated Approach of Ant Colony Optimization (ACO), Machine Learning (ML), and Fuzzy Logic for Revolutionizing Inventory Management in Modern Supply Chains

“…The fusion of ACO, ML, and Fuzzy Logic in inventory management aligns seamlessly with the principles of information sharing and interoperability within supply chains, echoing the sentiments of Khan and Abonyi (2022), particularly in the context of the burgeoning circular economy. Amidst a world where computational intelligence and optimization techniques hold increasing prominence, this integrated approach resonates with the trend of harnessing machine learning for solving combinatorial optimization problems, as discerned from the work of Yang et al (2022). This synthesis harmonizes with comprehensive efforts geared towards sustainably optimizing manufacturing processes through computational intelligence, as illuminated in "Computational Intelligence based Optimization of Manufacturing Process for Sustainable Materials" (Sinwar et al, 2023).…”

State-Space Compression for Efficient Policy Learning in Crude Oil Scheduling

“…This methodology models business challenges as Markov decision processes and learns policies that maximize cumulative rewards through sustained interaction with the environment. Its core strengths lie in its neural-network-based approximation capabilities, rapid sequential decision making, and a degree of adaptability in addressing dynamic programming challenges [13]. Yet, when applied to actual industrial problems, these methods often grapple with expansive state spaces, extended training durations, and convergence difficulties [14], signaling the need for more efficient methods.…”