SCSC2003 Abstract S41604
Search Algorithm for Nonlinear Stochastic System – Bay of Biscay Scenario
Search Algorithm for Nonlinear Stochastic System – Bay of Biscay Scenario
Submitting Author: Ms. Subhashini Ganapathy
Abstract:
Continuously evolving system state, uncertainty, and qualitative constraints make logistics planning and control problems complex. Military models more often than not handle complex issues. Several military models of warfare have focused on generating a closed-form analytic solution. Purely closed-form analytic solutions have limited value in solving the dynamic search problem, which involves strategy adaptation and continuous re-planning. In this article, we investigate approaches to integrate human reasoning with heuristics and optimization techniques in analysis of these systems in the context of a simulated historical event – the Bay of Biscay Scenario.
One of the major research challenges is to develop a search model that would include the temporal dynamics of the system and minimize the cost function with multiple constraints. The objective is to define the state of the search area as a mathematical representation of discrete change in time. The proposed model focuses on the U-boat hunting problem during World War II. It was on the Bay of Biscay area, off the coast of France and Spain. The German U-boats operated from ports in occupied France, crossing the Bay of Biscay to gain access to the North Atlantic. The U-boat wolf packs would look for and attack allied convoys. The objective of the allied aircrafts is to locate the U-boats and destroy them, before they can inflict damage.
The Bay of Biscay Scenario has multiple objectives. The first goal is to maximize the number of U-boats destroyed. Second goal is to minimize the time it takes to locate the U-boats and to destroy them. Finally, the goal is to minimize the number of convoys lost in the mission and maximize the utilization of the aircrafts and the other allied resources. The problem constraints include the number of allied aircrafts available for the mission, the limited maintenance resources and service available to support sortie generation, the characteristics of the aircraft such as maximum speed, range of detection, altitude and their munitions carrying capability.
This problem is a dynamic path-planning problem with uncertainty and the environment presenting hostile threats. Our approach integrates simulation, optimization, and human computer integrated heuristic reasoning. This article overviews the historical events, highlights the complexity of the problem solving process, presents an architecture to support modeling and simulation and describes our solution method.
Back to SCSC2003 Abstracts