Solving multiple travelling officers problem with population-based optimization algorithms

The travelling officer problem (TOP) is a graph-based orienteering problem for modelling the patrolling routines of a parking officer monitoring an area. Recently, a spatiotemporal probabilistic model was built for TOP to estimate the leaving probability of parking cars, and relevant algorithms were...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Neural computing & applications 2020-08, Vol.32 (16), p.12033-12059
Hauptverfasser: Qin, Kyle K., Shao, Wei, Ren, Yongli, Chan, Jeffrey, Salim, Flora D.
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The travelling officer problem (TOP) is a graph-based orienteering problem for modelling the patrolling routines of a parking officer monitoring an area. Recently, a spatiotemporal probabilistic model was built for TOP to estimate the leaving probability of parking cars, and relevant algorithms were applied to search for the optimal path for a parking officer to maximize the collection of parking fines from cars in violation. However, there are often multiple parking officers on duty during business hours in the central business district, which provides us with the opportunities to introduce cooperation among officers for efficient car-parking management. The multiple travelling officers problem (MTOP) is a more complex problem than the TOP because multiple officers are involved simultaneously in paths construction. In this study, the MTOP is formulated and new components are established for solving the problem. One essential component called the leader-based random-keys encoding scheme (LERK) is developed for the representation of possible solutions. Then, cuckoo search (CS), genetic algorithm (GA) and particle swarm optimization (PSO) are implemented using the proposed components and compared with other state-of-the-art GA and PSO using other solution encoding schemes to solve MTOP. In addition, two greedy selection algorithms are adopted as baselines. The performance of the algorithms is evaluated with real parking sensors data and different metrics. The experimental results show that the performance of CS and GA using LERK is considerably improved in comparison with that of other implemented algorithms.
ISSN:0941-0643
1433-3058
DOI:10.1007/s00521-019-04237-2