Guest Editorial Welcome to this special issue on intelligent mobile agent systems. Intelligent mobile agent systems are one of the rapidly developing mechanisms that can help organize information and data on computer networks and even the Internet. Two primary features distinguish intelligent mobile agent systems: autonomy and movement. Thus they are able to move to data sources and function efficiently and effectively to meet and address their design objectives. While agent systems offer significant advantages there are several challenges to be overcome in the development of these systems. Agent system reliability under dramatically changing loads involving both executing agents and agent messaging can be difficult to address. Agent systems may at times have only a few agents and at other time may involve many thousands of agents simultaneously operating with desperate goals. This challenge is referred to as system scalability. In this issue, Deters addresses this challenge by proposing and describing a scalable agent framework called DICE (Dependable Information Agent Computing Environment) which is FIPA compliant. This system is based on an approach of using many secure runtime environments called cells. Deters also provides examples of how this framework can be actually deployed in practical agent systems. Another important challenge associated with mobility is that of security. This challenge includes three forms of security: protecting the agent against malicious hosts, protecting the host against malicious agents, and protecting agent communication against eavesdropping and alteration. Voung and Fu address the first of these challenges by describing a rich security model in conjunction with the Actigen agent system. The security model provides different levels of security depending upon previously established access criteria and permissions. A third challenge is to address agent system coordination and control. This challenge has to do with the autonomy eof agents and the behaviour of agent of the system as an organized coUection of independent executing objects. Smith et al. examine this challenge in the context of environment sensing and load balancing. Agents' use progressively sophisticated models to attempt to predict server loads and autonomously move to sites, which they perceive as optimal. Agent behavior and agent system behavior are described and contrasted. The final contribution in this issue examines the challenge of developing quality software under the severe restrietions of the agent paradigm. Benedicenti et al. identify key metrics to monitor agent software quality and show how high quality software agent development can proceed. Thus, this special issue of the ACR brings together articles that address the important challenges in intelligent mobile agent system development. We hope the reader finds this issue interesting and stimulating. Dr. Raman B. Paranjape University of Regina

/lp/association-for-computing-machinery/guest-editorial-yOachGfzJe