Copyright © 2007 The Institute of Electronics, Information and Communication Engineers
Regular Section -- Papers -- Distributed Cooperation and Agents |
An Efficient and Privacy-Aware Meeting Scheduling Scheme Using Common Computational Space
1 The author is with The Graduate University for Advanced Studies, Tokyo, 101–8430 Japan. E-mail: huda{at}grad.nii.ac.jp, 2 The authors are with the National Institute of Informatics, Tokyo, 101–8430 Japan.
 |
Abstract |
|---|
Privacy is increasingly viewed as a key concern in multi-agent based algorithms for Distributed Constraint Satisfaction Problems (DCSP) such as the Meeting Scheduling (MS) problem. Many algorithms aim for a global objective function and as a result, incur performance penalties in computational complexity and personal privacy. This paper describes a mobile agent-based scheduling scheme called Efficient and Privacy-aware Meeting Scheduling (EPMS), which results in a tradeoff among complexity, privacy, and global utility. It also introduces a privacy loss model for collaborative computation, multiple criteria for evaluating privacy in the MS problem, and a privacy measurement metric called the Min privacy metric. We have utilized a common computational space in EPMS for reducing the complexity and the privacy loss in the MS problem. The analytical results show that EPMS has a polynomial time computational complexity. In addition, simulation results show that the obtained global utility for scheduling multiple meetings with EPMS is close to the optimal level and the resulting privacy loss is less than for those in existing algorithms.
Key Words: algorithm, constraint, distributed computing, mobile agent, privacy
Manuscript received March 14, 2006. Manuscript revised August 2, 2006.