Olive Niyomubyeyi
Doctoral student
A Comparative Study of Four Metaheuristic Algorithms, AMOSA, MOABC, MSPSO, and NSGA-II for Evacuation Planning
Author
Summary, in English
Evacuation planning is an important activity in disaster management to reduce the effects of disasters on urban communities. It is regarded as a multi-objective optimization problem that involves conflicting spatial objectives and constraints in a decision-making process. Such problems are difficult to solve by traditional methods. However, metaheuristics methods have been shown to be proper solutions. Well-known classical metaheuristic algorithms—such as simulated annealing (SA), artificial bee colony (ABC), standard particle swarm optimization (SPSO), genetic algorithm (GA), and multi-objective versions of them—have been used in the spatial optimization domain. However, few types of research have applied these classical methods, and their performance has not always been well evaluated, specifically not on evacuation planning problems. This research applies the multi-objective versions of four classical metaheuristic algorithms (AMOSA, MOABC, NSGA-II, and MSPSO) on an urban evacuation problem in Rwanda in order to compare the performances of the four algorithms. The performances of the algorithms have been evaluated based on the effectiveness, efficiency, repeatability, and computational time of each algorithm. The results showed that in terms of effectiveness, AMOSA and MOABC achieve good quality solutions that satisfy the objective functions. NSGA-II and MSPSO showed third and fourth-best effectiveness. For efficiency, NSGA-II is the fastest algorithm in terms of execution time and convergence speed followed by AMOSA, MOABC, and MSPSO. AMOSA, MOABC, and MSPSO showed a high level of repeatability compared to NSGA-II. It seems that by modifying MOABC and increasing its effectiveness, it could be a proper algorithm for evacuation planning.
Department/s
- Dept of Physical Geography and Ecosystem Science
- MECW: The Middle East in the Contemporary World
- Centre for Geographical Information Systems (GIS Centre)
- Centre for Advanced Middle Eastern Studies (CMES)
Publishing year
2020-01-03
Language
English
Publication/Series
Algorithms
Volume
13
Issue
1
Document type
Journal article
Publisher
MDPI AG
Topic
- Computer Science
- Physical Geography
Keywords
- Geospatial Artificial Intelligence (GeoAI)
- Artificial Intelligence (AI)
Status
Published
ISBN/ISSN/Other
- ISSN: 1999-4893