SimISpace
SimISpace is an agent-based simulation model that is being developed to implement the main features of the I-Space as a conceptual framework. It is designed to serve as a general knowledge management engine that, through a user-friendly graphical interface, can be adapted to a wide range of knowledge-related applications.We start by creating domain-specific knowledge networks that agents are required to discover and exploit by investing either in individual learning processes – codification, abstraction, etc. – or in interactions with other agents. The two together make up a social learning process. SimISpace allows a user to specify different knowledge groups, agent groups, discovery conditions, knowledge spillover conditions, cost conditions and reward conditions.
Complex scenarios can be modeled by assigning distributions for each of the initial properties of agent and knowledge groups. More than 160 direct parameters, both general and action-specific, can be set for each of the agent and knowledge groups created.
During the simulation runs, agents compete and cooperate in performing three different kinds of actions:
- Managing knowledge (e.g. discovering, codifying & abstracting, protecting, exploiting, disposing, etc.)
- Exchanging knowledge (e.g. trading, licensing)
- Networking and being in motion (e.g. meeting, relocating, relaxing, exiting & entering)
A detailed description of SimISpace and its innovative features can be found in the following paper: Ihrig, M. & Abrahams, A. (2007). "Breaking new ground in simulating knowledge management processes: SimISpace2", in: Proceedings 21st European Conference on Modelling and Simulation, Ivan Zelinka, Zuzana Oplatková, Alessandra Orsoni (Eds.), Prague, Czech Republic.
