Routing Sales Territory by Solving a Multi-objective TSP Variant with Evolutionary Algorithms

The Traveling Salesman Problem is a classic NP-hard combinatorial problem. However, many real-world scenarios do not match the classic TSP modeling. Thus, this paper proposes generating vendors' routes in a sales territory by extending and combining different TSP variants. The proposal deals wi...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Proceedings - International Conference on Tools with Artificial Intelligence, TAI S. 109 - 116
Hauptverfasser: Menezes Sampaio, Savio, Dantas, Altino, Camilo-Junior, Celso G.
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 01.11.2019
Schlagworte:
AI Applications
Artificial intelligence
Combinatorial Optimization
Customer Importance
Economics
Evolutionary computation
Frequency
Genetic Algorithms
IVF/NSGA-II
Multi Objective Memetic Algorithm
Multi-Objective Problem
Planning and scheduling
Proposals
Routing
Sales Territory
Schedules
Scheduling
Scheduling and Allocation
Time-frequency analysis
Traveling Salesman Problem
Traveling salesman problems
Urban areas
ISSN:2375-0197
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The Traveling Salesman Problem is a classic NP-hard combinatorial problem. However, many real-world scenarios do not match the classic TSP modeling. Thus, this paper proposes generating vendors' routes in a sales territory by extending and combining different TSP variants. The proposal deals with many constraints such as prioritizing the most valuable customers, resource limits to visit all customers, time and day windows, multiple days horizon, lunch stop, visit frequency and destination importance. We also introduce a permutational representation with a series of directed arcs, and with a slot to indicate customers that could not be visited. We evaluate SPEA2, NSGA-II, and IVF/NSGA-II Multi-Objective Evolutionary Algorithms against three real-world scenarios. The results demonstrate that the proposed model can produce good results in a realistic scenario, maintaining a balance between prioritizing important customers and economic routes. Besides, IVF/NSGA-II outperformed the other algorithms in most cases, and also the manually-created routes, e.g., reducing 35% the route's distance and increasing 60% the route's importance.
ISSN:2375-0197
DOI:10.1109/ICTAI.2019.00024