![]() |
|||
![]()
|
![]() |
![]() |
![]() |
4.2. MOBILE AGENTSMobile agents travel on the computer network to go to the site of a server they need to use. The interaction with the server occurs at the server site, and then agents travel back to their origin to provide users the results. Essentially, mobile agents are a solution to the limited bandwidth of the network. By sending an agent to the server site, the amount of network traffic generated is minimized. For example, consider a library catalog server on the network. A nonmobile interface agent queries the server for certain items. The results of the query are returned to the agent. Suppose the desired item is not available at this server and the agent must query another server. In this case, the network transmission from the first server can be avoided if the agent were to visit the server site. Then, upon finding that the server does not have the desired item, the agent could travel to the second server site, and so on. Mobile agents are particularly attractive for mobile workers. Typically, the network bandwidth available from an offsite location is quite limited. In this case, agents that travel on the network and return with results are an answer to the limited bandwidth. One of the difficult issues with mobile agents is security. Methods to verify the trustworthiness of arriving agents at a server are needed prior to allowing them access to data and computing resources. With regard to computing resource consumption, mobile agents should be self-aware of their resource consumption relative to the server load to avoid overwhelming the server. A mobile agent must also provide a means to securely carry gathered data back to the user. 4.3. AUTONOMOUS AGENTSAutonomous agents are computational systems designed to operate in a changing, unpredictable universe. Within their universe these systems sense their environment through sensory inputs, and they may explore their surroundings through the execution of commands that effect change either literally for robotic systems, or figuratively for those systems that wander through cyberspace. The problem of autonomous systems is to understand how they develop and modify the principles by which they control their behavior while becoming and remaining effective as systems achieving their goals in complex dynamic environments. Autonomous systems are those that develop, for themselves, the laws and strategies according to which they regulate their behavior. Once started, they determine their actions on their own. The term "autonomous agents" can refer to either to roving robotic systems of various forms, or to purely software robots or softbots that have no physical form, and explore the universe of computer memory and computer networks. These agents may perform a variety of tasks in their environments. For example, some robots may work in hazardous physical environments, interacting primarily with other robots and physical objects. On the other hand, softbots may act as personal assistants to office workers by running on a single PC or workstation, or may serve a wider set of users by running on a network. Such agents use knowledge about the interests and priorities of people to perform routine organizational tasks such as automatically screening, directing, revising, and responding to information. With the explosion of interest in the Internet, there are a number of software agents currently available or under development to aid in the search, filtering, and organization of information on the World Wide Web. Such systems are often called searchbots. Agents that modify their behavior over time to adapt to their environment are viewed as intelligent autonomous agents. Research in this area is concerned with developing autonomous agents that can improve their performance over time, or "learn" to adapt to their changing environment. 4.4. COOPERATION AND COORDINATIONAs autonomous agents are developed to take over a variety of tasks for people, there is a need for developing agents that can cooperate and coordinate with other agents operating in the same environment. (Have your agent get in touch with my agent to arrange a meeting.) Developing systems so that agents can communicate among themselves as well as sensing and operating within their environment is the focus of research in Distributed Artificial Intelligence. Such systems may either consist of multiple agents working together to solve a single problem; or multiple agents, each working to solve their own problem, that occasionally need to interact with other agents to achieve their goals. Research under the direction of Professor Victor R. Lesser at the University of Massachusetts at Amherst focuses on how to achieve effective cooperation among agents, balancing several interdependent criteria including efficient use of processor and communication resources, responsiveness to unexpected situations and real-time deadlines, and reliability. Three conceptual ideas motivate Lesser's approach to achieving effective coordination: satisficing versus optimality, tolerance of inconsistency, and control intelligence. Agents solving independent goals may exploit each other's different sets of world knowledge. Instead of each agent relearning what other agents have already learned through experience, agents can simply ask each other for help. This gives each agent access to the experience that other agents already possess. This is the focus of the work of Pattie Maes et al. at the MIT Media Laboratory. For agents to communicate and collaborate, they must speak a common language as well as follow a common protocol, as discussed in Section 4.1.
|
![]() |
|
Use of this site is subject certain Terms & Conditions. Copyright (c) 1996-1999 EarthWeb, Inc.. All rights reserved. Reproduction in whole or in part in any form or medium without express written permission of EarthWeb is prohibited. Please read our privacy policy for details. |