11. Conclusion and Future Work

Abstract

Inspired by constraint technology and agent technology, we have in this book engaged in a comprehensive study of collaborative problem solving, its advantages and pitfalls, but mainly its improvement through autonomous dynamic reconfiguration. Autonomous dynamic reconfiguration is a complete technology dealing with two of the most common problems in collaborative problem solving — namely poor problem solving results and low efficiency due to high communication efforts. This technology includes a profound formal analysis of the problems in collaborative problem solving, a necessary and sufficient set of practical concepts and techniques on the micro- and macrolevel of agency for a more flexible collaborative problem solving process, and an integrative control approach to autonomously and dynamically adapt the problem solving system to the current problem structure.

From the beginning on, this work has dealt with the configuration of a collaborative problem solving effort, i.e. the distribution of problem solving knowledge, goals and skills. The main idea has been the introduction of two individual operations to locally change this configuration — agent melting and agent splitting. Agent melting means to unify the problem solving knowledge, goals and skills of two or more agents in a single agent, while agent splitting denotes a process in which a single agent splits its problem solving knowledge, goals and skills and hands it over to one or more new or existing agents.