Assigning weapons to targets is a key aspect of operational planning in wargames. In past wargames, these assignments were built by subject matter experts with a deep understanding of each weapon system, platform, and target. However, the time-consuming process of manually building an Integrated Tasking Order (ITO) creates a strain on the game-move process. To adapt to increasing demand for speed and effectiveness, A.I. methods to automate the assignments must be explored. It is a classic example of the Weapon Target Assignment Problem (WTAP[BM[1] ) in which a set of various weapons is assigned to a set of targets to maximize damage to the enemy. The WTAP can be modeled and defined using tactical and logistical constraints, including platforms’ and munitions’ compatibility, availability, lethality, and range. This paper describes a customized genetic algorithm (GA) and demonstrates its ability to solve for a feasible weapon-to-target operational plan. Due to the vast solution space of all possible assignments, finding an optimal solution is computationally expensive, but in a fast-paced wargaming environment, the need for rapid results is imperative. Empirical testing of this genetic algorithm on the problem proves its capability to obtain an effective ITO quickly. We simulate the attacks planned in the ITO using Monte Carlo and analyze the performance by comparing the attack results at various levels of GA convergence. Furthermore, optimizing ITOs in each successive move leads to efficient utilization of in-game resources and more damage to targets over the course of the game.

Keywords
AI, SIMULATIONS

Additional Keywords
genetic algorithm, optimization, Weapon Target Assignment Problem, wargames