Abstract
Agent Oriented Engineering (AOE) represents an exciting new means of analyzing and designing complex software systems. Mobile agents have proved as an effective solution in building wide range of computing distributed applications. Despite the research papers carried out, a number of disciplines concerning mainly performance optimization of mobile agent-based applications are still growing. In this paper, we have successfully developed an approach that computes performance optimization of mobile agent systems. This approach utilizes Master-Slave Design Pattern of mobile agent systems. The proposed approach is called the Least Time performance optimization which uses a new developed computing model, namely Message Broadcast Model (MBM). Basically, MBM manages messages between set of master agents carrying out tasks on a set of clients and corresponding slave agents receiving and performing tasks on servers. Each slave agent registers its working time given in milliseconds in message broadcast handler. We analyzed MBM using new types of matrices namely Message Delivery/Response Matrix (MDRM). Performance optimization is calculated based on Least Time approach and compared with another approach namely V-agent. Furthermore, we implemented a prototyping scenario of MDRM based on MBM. This approach is developed as a mean of exploiting analyses for parallel distributing computing. Keywords: Mobile Software Agents, Message Broadcast Model (MBM), Message Delivery/response Matrix (MDRM), Least Time Approach, Performance Analysis.