Home » A Tabu Search Approach to the Weapons Assignment Model by Christopher A Cullenbine
A Tabu Search Approach to the Weapons Assignment Model Christopher A Cullenbine

A Tabu Search Approach to the Weapons Assignment Model

Christopher A Cullenbine

Published September 17th 2012
ISBN : 9781249400837
Paperback
106 pages
Enter the sum

 About the Book 

USSTRATCOM/J535 developed the Weapon Assignment Model (WAM) to analyze potential future force structures and build the Red Integrated Strategic Offensive Plan (RISOP), a hypothetical Russian war plan for attacking the West. Typical problems from WAMMoreUSSTRATCOM/J535 developed the Weapon Assignment Model (WAM) to analyze potential future force structures and build the Red Integrated Strategic Offensive Plan (RISOP), a hypothetical Russian war plan for attacking the West. Typical problems from WAM consider 3,000 targets and 6,000 warheads from about 2,700 weapon systems. The Multiple Independent Reentry Vehicle (MIRV) systems are limited to a pool of about 30,000 feasible footprints. The integer linear formulation of this problem often has over 16 goals, 500,000 decision variables, and 10,000 constraints. The problem size warrants the use of heuristics to find good solutions. This research demonstrates that tabu search obtains excellent solutions to problems from WAM. The tabu search basically deletes and adds individual weapon assignments to find better solutions, while forbidding (tabu) backtracking on a move for a number of iterations (tabu tenure). Very good solutions are obtained in about 30 minutes on a dual processor PC when starting with no weapons assigned. The tabu search obtains good solutions much faster if a previous solution, even if currently infeasible, is the starting point for the search. While significantly reducing computational time from the current solve technique, the tabu search allows for either prioritized or goal programming, does not limit the size of the footprint pool, and is not sensitive to the order of the goals. Tabu searchs ability to quickly solve these nuclear targeting problems provides better insight into force structure and policy decisions. This research lays the foundation for automating and optimizing U.S. nuclear war plans.