Hybrid Architecture supporting Multiple Models of Enemy Reactions (HAMMER)

Commanders engaged in modern counterinsurgency warfare achieve success by accounting for adversary behaviors and responses to courses of action during the planning process.  A single type of modeling formalism is unable to capture all of the dynamics of modern asymmetric warfarethere are too many levels and types of analysis for a single modeling paradigm to be sufficiently valid and accurate across all possible operational environments and planning activities.  To effectively explore the potential reactions of an adversary, we need to incorporate relevant models of the enemys strategies and tactics across the resource allocation, planning and scheduling, and mission execution spaces.  Planners also need to know what information is necessary to confirm, refute, or refine forecasts.  Our plan is to develop HAMMER, a hybrid architecture that incorporates meta-models to represent each models technical and data requirements (including inputs and outputs), theoretical underpinnings, and operational uses.  The meta-models will enable the planner to decompose forecasting tasks into subtasks solvable by other models; they will facilitate the input of data from multiple repositories or data mining tools; and they will allow the planner to know when additional data collection could confirm, refute, or refine forecasts. BENEFIT: We expect the full-scope HAMMER system to have immediate and tangible benefit for political decision makers, Commanders in all branches of the U.S. military, and the intelligence community. Adding support for more thorough validation of models and to better provide for the end-users decision-making needs provides additional beneficial to potential customers of our commercial AgentWorks product, such as those using complex, agent-based models to support financial and health-management decisions.