Multi-Agent Geosimulation In Support To “What if” Courses Of Action Analysis

Abstract

In this paper we propose an approach to support "what-if" analysis in the context of COA evaluation. Our approach consists in using multiagent geosimulation to simulate the execution of COAs in a Virtual Geographic Environment (VGE) which can change during the simulation, and then allowing the user to explore various assumptions and to analyse their outcomes. We identify the requirements to support this approach and we present how we implement them in the MAGS-COA software. We also illustrate our approach on an example and we present future works.

Sociocultural Behavior Sensemaking: State of the Art in Understanding the Operational Environment

“…Research by Laval University in Canada has also explored explicit use of Multi-Agent Geosimulation (MAGS) to support qualitative analysis of different COAs (Haddad & Moulin, 2008). This system is designed specifically to facilitate the type of "what if" analysis necessary for comparing the outcomes and side effects of various COAs in a complex sociocultural environment.…”

“…Users can explore various COAs and assumptions through the agent-based simulations, modeling how the environment evolves over time as the COA is implemented. The MAGS-COA system also directly facilitates COA comparison, producing a qualitative report at the end of the simulation runs that documents the impacts of various potential COAs and their relative success in achieving particular mission goals (Haddad & Moulin, 2008). Similar analyses can also be run using the multi-agent understanding and forecasting simulations described in Chapters 2 and 10.…”