Planning of complex supply chains: A performance comparison of three meta-heuristic algorithms

- A mixed-integer nonlinear model for tactical planning of a green supply chain.
- Solving the model using Genetic Algorithm, Simulated Annealing and Cross-Entropy.
- Utilizing data from an Australian clothing manufacturer.
- Comparative analyses of the numerical results.
- We find that Cross-Entropy method outperforms the two popular meta-heuristics.
- Simulated Annealing can produce better results in a time-restricted comparison.

Businesses have more complex supply chains than ever before. Many supply chain planning efforts result in sizable and often nonlinear optimization problems that are difficult to solve using standard solution methods. Meta-heuristic and heuristic solution methods have been developed and applied to tackle such modeling complexities. This paper aims to compare and analyze the performance of three meta-heuristic algorithms in solving a nonlinear green supply chain planning problem. A tactical planning model is presented that aims to balance the economic and emissions performance of the supply chain. Utilizing data from an Australian clothing manufacturer, three meta-heuristic algorithms including Genetic Algorithm, Simulated Annealing and Cross-Entropy are adopted to find solutions to this problem. Discussions on the key characteristics of these algorithms and comparative analysis of the numerical results provide some modeling insights and practical implications. In particular, we find that (1) a Cross-Entropy method outperforms the two popular meta-heuristic algorithms in both computation time and solution quality, and (2) Simulated Annealing may produce better results in a time-restricted comparison due to its rapid initial convergence speed.