Parallel evolutionary programming techniques for strategy optimisation in air combat scenarios

Nusyirwan, I 2008, Parallel evolutionary programming techniques for strategy optimisation in air combat scenarios, Doctor of Philosophy (PhD), Aerospace, Mechanical and Manufacturing Engineering, RMIT University.

Document type: Thesis
Collection: Theses

Attached Files
Name Description MIMEType Size Attachments (.avi files) Click to show the corresponding preview/stream application/zip 22.13MB
Nusyirwan.pdf Thesis application/pdf 4.92MB
Title Parallel evolutionary programming techniques for strategy optimisation in air combat scenarios
Author(s) Nusyirwan, I
Year 2008
Abstract Air combat between fighter missiles and aircraft can be categorised as a pursuit-evasion problem. One aircraft acts as a pursuer and the other as an evader. Generally, the pursuer will try to capture the evader as quickly as possible and the evader tries to evade capture for as long as possible. For an experienced human pilot, it is trivial to discuss this methodology, but to simulate it, the mathematics involved is very complex and difficult to implement in a computer environment. Classical methods, though very accurate in their analysis, are not suited to solve a complex 6DOF pursuit-evasion problem and they have limitations in representing real-world problems such as discontinuities, discrete, stochastic, chaotic, temporal information or lack of information. In this thesis, evolutionary programming (EP) is applied to determine the optimum maneuvering strategy for an aircraft (evader) to avoid interception by an incoming missile (pursuer). EP is a class of algorithms known as Evolutionary Algorithm (EA). EA has an ability to find an optimal solution in a complex problem which involves discontinuities, discrete, nondifferential parameters and noise. In addition, the methodology was implemented on parallel computer architecture to improve the computing time and expanding the search space. A sensitivity analysis was carried out to determine the best configuration and to understand the effect of parameters, such as number of processors, population size, number of generations, etc., on the results. The effects of sensor and instrument errors were also considered. The method enabled feasible solutions to be found in a relatively short period of time. However, the ability to search for feasible solutions is dependent on various parameters such as initial conditions, aircraft configurations and aerodynamic constraints. It is concluded that, in general, EP is able to determine feasible maneuvering strategies for an evader to avoid interception with and without instrument errors. The methodology has the potential to be used as a training tool for pilots in air combat or as an intelligent engagement strategy for autonomous systems, such as Unmanned Air Combat Vehicles (UCAV).
Degree Doctor of Philosophy (PhD)
Institution RMIT University
School, Department or Centre Aerospace, Mechanical and Manufacturing Engineering
Keyword(s) Parallel evolutionary programming
game theory
Version Filter Type
Access Statistics: 376 Abstract Views, 830 File Downloads  -  Detailed Statistics
Created: Tue, 30 Nov 2010, 14:47:22 EST by Guy Aron
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us