Multiple Hungarian Method for k-Assignment Problem

• Published in: Mathematics (Basel) Vol. 8; no. 11; p. 2050
• Main Authors: Gabrovšek, Boštjan, Novak, Tina, Povh, Janez, Rupnik Poklukar, Darja, Žerovnik, Janez
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01.11.2020
• Subjects: Computation; Datasets; Genetic algorithms; greedy algorithm; Heuristic; heuristic algorithm; hungarian method; k-assignment problem; k-matching problem; Linear programming; local search; Mathematics; Operations research; Optimization; Production planning; Search algorithms
• ISSN: 2227-7390; 2227-7390
• DOI: 10.3390/math8112050

The k-assignment problem (or, the k-matching problem) on k-partite graphs is an NP-hard problem for k≥3. In this paper we introduce five new heuristics. Two algorithms, Bm and Cm, arise as natural improvements of Algorithm Am from (He et al., in: Graph Algorithms And Applications 2, World Scientific, 2004). The other three algorithms, Dm, Em, and Fm, incorporate randomization. Algorithm Dm can be considered as a greedy version of Bm, whereas Em and Fm are versions of local search algorithm, specialized for the k-matching problem. The algorithms are implemented in Python and are run on three datasets. On the datasets available, all the algorithms clearly outperform Algorithm Am in terms of solution quality. On the first dataset with known optimal values the average relative error ranges from 1.47% over optimum (algorithm Am) to 0.08% over optimum (algorithm Em). On the second dataset with known optimal values the average relative error ranges from 4.41% over optimum (algorithm Am) to 0.45% over optimum (algorithm Fm). Better quality of solutions demands higher computation times, thus the new algorithms provide a good compromise between quality of solutions and computation time.