An improved algorithm for deadlock detection and resolution in mobile agent systems

Mobile agent systems have been proved that are the best paradigm for distributed applications. They have potential advantages to provide a convenient, efficient and high performance distributed applications. Many solutions for problems in distributed systems such as deadlock detection rely on assumptions such as data location and message passing mechanism and static network topology that could not be applied for mobile agent systems. In this paper an improved distributed deadlock detection and resolution algorithm is proposed. The algorithm is based on Ashfield et. al. process. There are some cases in which original algorithm detects false deadlock or does not detect global deadlocks. The proposed algorithm eliminates the original algorithm deficiencies and improves its performance. It also minimizes the detection agent travels through the communication network. Also it has a major impact on improving performance of the mobile agent systems.