Decision Support for the Weapons-Target Assignment Problem under Uncertainty

Kathlyn Low Ying Hui

Abstract

In an age where speed and uncertainty dominate the battlefield environment, the need for faster and even more robust decision tools to ensure the optimal assignment of weapons to targets for achieving effective engagements during a battle, cannot be further emphasized. Absence of such tools is likely to result in undesirable consequences, for example, the suffering of hefty monetary losses by decision makers.

Recent studies on such assignment problem have begun to take into account the stochastic and uncertain behavior of parameters in the problem. In addition, they have also introduced the methodology of using risk measures to manage risks that arose as a result of uncertainty in the distributions of these parameters. However, a major issue exists in that, these studies require readers to have a sound knowledge and thorough understanding of risk measures before they are able to use the formulations provided to solve their problems. Such a requirement has certainly posed much inconvenience and hindrance to military decision makers who lacks the required knowledge.

In recognition of this problem, this thesis has proposed a system approach based on a new formulation for the Weapon-Target Assignment problem. Instead of adopting the approach proposed by existing literature, which is to use complex risk measures to manage risks and losses, this new formulation is based on a more easily understood approach the probability approach. It has thus, removed the need for readers to have domain knowledge in the area of risk measures before being able to employ this tool.

Other areas covered by this thesis includes: (1) the analyses of numerical results obtained from both previous and the newly proposed formulations of the Weapon- Target Assignment problem, (2) suggestions on how improvements can be made to limitations identified in previous formulations, (3) provision of algorithms for generating efficient frontiers, and lastly, (4) presentation of outcomes of sensitivity analysis studies conducted.