Using Cooperative Multi-agent Systems to Support the Generation of Student Learning Profiles in the Deep Dive Land Bridge System

The goal of the project described here was to extend the immersive experience of the user in the Virtual Land Bridge. Such extensions would facilitate the use of the software system to contribute to the development of individual student learning plans over several different age categories. The principal strategy was to modify the multi-agent planning system to allow for the generation of local tactical behavior of the agents in the virtual world. This resulted in the addition of a new layer to the MAS planning system, producing a hybrid system that utilized global knowledge monolithically in the Pathfinder portion and local tactical knowledge in the VR layer. The resultant hybrid system was designed to produce improved caribou agent decision making and movement that could scale up to support large herds on the order of a hundred thousand and more. This new system is shown to increase user immersion that can contribute to the development of student learning profiles.