Authors:
Himadri Khargharia
;
Sid Shakya
and
Dymitr Ruta
Affiliation:
EBTIC, Khalifa University, Abu Dhabi, U.A.E.
Keyword(s):
Trade Data Harmonisation, Genetic Algorithm, Population-Based Incremental Learning, Distribution Estimation Using MRF, Simulated Annealing.
Abstract:
The harmonization of trade data from two datasets containing different and distinct categories poses a challenging real-world problem. To address this issue, we model it as an optimization problem and investigate the effectiveness of various metaheuristic techniques in achieving optimal or near-optimal solutions. Particularly, we analyze the performance of Genetic Algorithm (GA), Population-based Incremental Learning (PBIL), DEUM, and Simulated Annealing (SA) in terms of best fitness, scalability, and their respective strengths and weaknesses. We explore multiple instances of the trade data harmonisation problem of different sizes to assess the applicability of these techniques in mitigating trade volume disparities. By examining the outcomes, our research offers valuable insights into the suitability of metaheuristic techniques for this problem.