Abstract

The purpose of this paper is to study how a multi-agent-based coordination mechanism can resolve conflicts among functional units in an enterprise. In the era of information technology, characterized by web technology, intelligent agents demonstrate the best interests of a functional unit when engaged in the highly delicate coordination process of an enterprise. Thus, behavioral problems that occur while coordinating conflicts among functional units can be avoided effectively. Since there are a number of functional units in an enterprise, the multi-agents are utilized to support the agent-based coordination of conflicts among several functional units. This paper proposes a new type of multi-agent-based coordination engine termed MACE in which the information exchange between multi-agents is properly controlled via a complicated algorithm. To illustrate the performance of MACE, we devised a cooperative organizational decision support system (CODSS) and incorporated MACE into CODSS, yielding MACE_CODSS. The combined case of production and marketing was introduced and tackled by MACE_CODSS, obtaining the proper coordination results. To verify the improved performance of MACE, we implemented three types of coordination DSS: Joint_CODSS, Separate_CODSS, and None_CODSS. With 12 simulation periods and 100 simulation runs, we obtained comparative coordination results for each coordination type. Statistical tests revealed that the proposed MACE_CODSS produced the most desirable coordination results.

Author Keywords: Multi-agent coordination; Cooperative organizational decision support systems; Production/marketing coordination; Marketing dominating coordination approach; Production dominating coordination approach; Joint coordination; Separate coordination

Article Outline

1. Introduction

2. Intelligent agents

2.1. Fundamentals

2.2. Multi-agents and coordination

3. Coordination and conflict resolution

3.1. Fundamental concept

3.2. Coordinated production/marketing decision-making

4. CODSS

4.1. Design perspective

4.2. MACE

4.2.1. MDCA and PDCA: control activity

4.2.2. Pseudo-code

5. Experiments

6. Concluding remarks

References

Corresponding author; email: kob@handong.edu