Genetic Algorithms Applied to a Mission Routing Problem
Abstract
This thesis applies genetic algorithms to a mission routing problem. The mission routing problem involves determining an aircraft's best route between a staging base and a target. The goal is to minimize the route distance and the exposure to radar. Potential routes are mapped to a 3-dimensional mesh where the nodes correspond to checkpoints in the route and the arcs correspond to partial paths of the route. Each arc is weighted with respect to distance and exposure to radar. A genetic algorithm is a probabilistic search technique loosely based on theories of biological evolution. Genetic crossover and survival of the fittest are the basis of a genetic algorithm's control structure and can be used for general problems. Encoding and evaluation of the data structure is problem specific. This thesis focuses on approaches to mapping the mission routing problem's mesh to a data structure which the genetic algorithm's control structure can effectively and efficiently manipulate. Three broad methods are proposed: Bounded Progress, Free Progress, and Restricted Progress. The Bounded Progress method uses a tightly-coupled mesh while the Free Progress method uses a loosely-coupled mesh. The Restricted Progress method is a hybrid of the other two methods.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1993
- Accession Number
- ADA274130
Entities
People
- James B. Olsan
Organizations
- Air Force Institute of Technology