A Comparison between JACK Intelligent Agents and JACK Teams Applied in Teamwork

Abstract

This report investigates JACK Intelligent Agents and JACK Teams, and makes a comparison between the two. The main object was to find indications that point out which modeling paradigm that results in least development effort, and which one that is creating the most feasible platform regarding teamwork construction. The application domain is decision-support systems used in oil production. The aspects evaluated are development effort, degree of coupling, encapsulation of functionality, abstraction level, delegation of autonomy, and scalability. The solutions developed in the comparison had static team formations that included few teammembers. This caused less development effort by using JACK Intelligent Agents, and was the main reason why it was considerate to be the preferred modeling paradigm in this case. This was partly experienced because reasoning based on the actual team membership was not used in the JACK Teams version. The use of roles was used instead, causing more JACK entities where it could have been avoided. Dynamic team formations during runtime were not needed due to the reference problem introduced. Maintaince during runtime, for instance introducing new subteams and changing the role structure was not looked into. Introducing teams in large scale was not performed. These four factors could have caused a different result. The question is if JACK Teams shows its potential through the oil production system designed in this report.