Using a genetic algorithm to solve the troops-to-tasks problem in military operations planning. (October 2017)

• Record Type: Journal Article
• Authors: Fauske, Maria Fleischer
• Is Part Of: Journal of defense modeling and simulation. Volume 14:Number 4(2017:Oct.)
• Page Start: 439
• Page End: 446
• Publication Date: 2017-10
• Subjects: Scheduling -- genetic algorithms -- military applications; Military art and science -- Computer simulation -- Periodicals; 355.0011305
• DOI: 10.1177/1548512917711310
• Languages: English
• ISSNs: 1548-5129

Abstract

The troops-to-tasks analysis in military operational planning is the process where the military staff investigates who should do what, where, and when in the operation. In this paper, we describe a genetic algorithm for solving troops-to-tasks problems, which are typically solved manually. The study was motivated by a request from Norwegian military staff, who acknowledged the potential for solving the troops-to-tasks analysis more effectively by using optimization techniques. Also, NATO's operational planning tool, TOPFAS, lacks an optimization module for the troops-to-tasks analysis. The troops-to-tasks problem generalizes the well-known resource-constrained project scheduling problem, and thus it is very difficult to solve. As the troops-to-tasks problem is particularly complex, the main purpose of our study was to develop an algorithm capable of solving real-sized problem instances. We developed a genetic algorithm with new features, which were crucial to finding good solutions. We tested the algorithm on two different data sets representing high-intensity military operations. We compared the performance of the algorithm to that of a mixed integer linear program solved by CPLEX. In contrast to CPLEX, the algorithm found feasible solutions within an acceptable time frame for all instances.