An algorithm to estimate enemy's location in WarGame based on pheromone

In the War-Game, the enemy's location is the key information for decision-making, but it's hard to be completely known. And the lack of current location information is one of the typical problems in incomplete-information decision. This paper proposes a pheromone-based model and relevant method for enemy's location estimation in an army tactics War-Game system3. In this paper, the key decision-making factors for moving a unit to a position are modeled as 8 kinds of pheromone; and the unit's intention is modeled as an 8-deimension vector, which represents the attracting or repelling degree for each pheromone. The join effects of all pheromone is taken as the attraction of each location. The location with the most attraction is estimated as the most possible result. The experimental results on the dataset of first national War-Game match show the proposed algorithm is feasible and practical: the top 3 estimation for the enemy who has historical observation is more than 70% accuracy. Moreover, the estimation with one step offset is over 90% accuracy, which is always can be tolerated in War-Game.