Introduction to Intelligent Software Agents Martin Beer, School of Computing & Management Sciences, Sheffield Hallam University, Sheffield, United Kingdom

Contents Agents as Tools of the Information Society –Information Society –Intelligent Software Agent as a tool of Information Society –Benefits from Intelligent Agent Fundamental Concepts of Intelligent Software Agents –What is an Agent? –Definition of Intelligent Software Agents –Characteristics of Intelligent Software Agents –Classification

The Information Society On the Way to the Information Society –Digitization –Networking –Internet Tools of the Information Society –Browsers –Search Engines

Intelligent Software Agents Intelligent Agents are –a new category of information society tools –software programs that independently perform tasks on behalf of a user in a network environment Intelligent Agents in the Business Area Intelligent Agents in the Private Area

Economic Potential(1/2) Potential Benefit from the User’s Viewpoint –Improvement of efficiency in working with internet Time-saving Increase in speed of arriving at solutions in internet –Improvement of effectiveness in working with internet Overcome the characteristics of human problem solving –Increase in the transparency and optimizations Comparing information from various sources User can select the most favorable Intelligent agents have made Adam Smith’s ‘Invisible hand’ visible in the electronic business world

Economic Potential(2/2) Potential Benefit from the Viewpoint of the Information and Communications Industry –Sales for intelligent agent (OVUM’s expectation) 19 million dollars in million dollars in billion dollars in 2006 –OVUM expects that the agents for private household will become the primary innovation factor and from the start of the next millennium the sales with private households will exceed those of business and government

What is an Agent? Agents are autonomous –Capable of acting independently –Over their own internal state System Environment Input Output

What makes Agents Interesting? Trivial (ie non-interesting) agents –Thermostat –Unix daemon (biff, ftpd etc) An intelligent Agent is a computer system capable of flexible autonomous action in some environment (Mike Wooldridge) By flexible we mean –Reactive –Pro-active –social

Reactivity If a program’s environment is guaranteed to be fixed, the program need never worry about its own success or failure –Example of a fixed environment – a compiler The real world is not like that Things change Information is incomplete –Many (most?) environments are dynamic

Reactivity Software is hard to build for dynamic environments –Program must take into account the possibility of failure –Ask itself whether it is worth executing! A reactive system is one that maintains an ongoing interaction with its environment and responds to changes that occur in it –(in time for the response to be useful)

Proactiveness Reacting to the environment is easy –(eg stimulus -> response rules) But we generally want agents to do things for us Hence goal directed behaviour Proactiveness = generating and attempting to achieve goals –Not driven solely by events –Taking the initiative Recognising opportunities

Social Ability The real world is a multi-agent environment –We cannot go around attempting to achieve goals without taking others into account Some goals can only be achieved with the cooperation of others Internet –Similarly for many computer environments – witness the Internet Social ability in agents is the ability to interact with other agents (and possibly humans) via some kind of agent communication language and perhaps cooperate with others

Other Properties Mobility –The ability of an agent to move around an electronic network Veracity –An agent will not knowingly communicate false information Benevolence –Agents do not have conflicting goals, and that every agent will therefore always try to do what it is asked of it Rationality –Agent will act to achieve its goals, and will not act in such a way as to prevent its goals from being achieved – at least insofar as its beliefs permit Learning/adaption –Agents improve their performance over time

Definition of Intelligent Software Agents(1/2) Three Major Categories of Agents –Human Agent / Hardware Agent / Software Agent

Two Major Requirement for Intelligent Agent –Autonomous Processing –Communication / Cooperation with other objects A Definition of Intelligent Software Agent is –a software program that can perform specific tasks for a user and possesses a degree of intelligence that permits it to perform parts of its tasks autonomously and to interact with its environment in a useful manner –Information / Cooperation / Transaction Agent Definition of Intelligent Software Agents(2/2)

Information Agent Task –search for information in distributed systems or networks locate information sources extract information from the sources filter the information of relevance for the user from the total quantity of found information using the user’s interest profile prepare and present the results in an appropriate form

Cooperative Agents Task –solve complex problems by using communication and cooperation mechanisms with other objects, such as agents, humans or external resources used when a problem exceeds the capabilities of an individual agent or agents already exist that already have a solution and whose knowledge can be used by other agents The demands on intelligence are higher than that of pure information agent

Transaction Agent Task –processing and monitoring of transaction in classical database environments and in the areas of network management and electronic commerce Such agents normally operate in very sensitive areas and represent their users for tasks that demand a high degree of responsibility, for example in the purchase of products using a user’s credit card

Characteristics of Intelligent Software Agents(1/2) The characteristics can be grouped into two large categories : internal and external Internal properties –determine the actions within the agent –Ex : the ability to learn, reactivity, autonomy, goal-oriented External properties –affect the interaction of several agents or human- agent communication –Ex : communication, cooperation

Characteristics of Intelligent Software Agents(2/2)

Internal Properties(1/4) Reactivity : –Capability of reacting appropriately to influences or information from its environment –Agent must have suitable sensors or possess its own internal model of its environment Proactivity / goal-orientation –Capability for an agent itself to take the initiative under specific circumstances –Require that the agent has well-defined goals or even a complex goal system

Internal Properties(2/4) Reasoning / learning –The intelligence of an agent is formed from three main components: its internal knowledge base, the reasoning capabilities based on the contents of the knowledge base, and the ability to learn or to adapt to changes to the environment. –Reasoning power requires rationality –Rational processing requires the existence of a goal system –Ability to learn from previous experiences and to successively adapts its behavior to the environment is as important for the intelligent behavior of an agent

Internal Properties(3/4) Autonomy –Capability to follow its goals autonomously, that is, without interactions or commands form the environment –To permit autonomous behavior, agent must be provided with those resources and capabilities, e.g. availability of an electronic network, capability to navigate through the network(mobility), capability to make contact with other agents(communication). Goal-orientation and ability to learn.

Internal Properties(4/4) Mobility –Ability to navigate within electronic communications networks –Demands on network environment and raise questions regarding security, data privacy, and management –Reduces the network loading. It can go to the computer or agents with the required information, not sending a series of messages over network. Then perform all tasks locally on the target computer –Agency : Meeting points, serves as marketplace or discussion and communication forum

External Properties(1/2) Communication / Cooperation –Interaction with its environment(human users, other agents, arbitrary information sources) –Communication require standardized protocol for exchange of information –Cooperation must augment the communications capability

External Properties(2/2) Character –It’s desirable for agent to demonstrate human traits –An agent’s most important characteristics are honesty, trustworthiness, and reliability –It is important that agents with a high degree of interaction with people exhibit emotional states, such as joy, sadness, frustrations.

Classification Three Criteria –Intelligence, Mobility, Number of Agents

Information Agents

Cooperation Agents

Transaction Agents

Agents and Objects Are agents just objects by another name? Object –Encapsulates some state –Communicates by message passing –Has methods, corresponding to operations that may be performed on this state

Main Differences Agents are autonomous –Agents embody a stronger notion of autonomy than objects, and in particular, they decide for themselves whether or not to perform an action on request from another agent Agents are smart –Capable of flexible (reactive, pro-active, social) behaviour, and the standard object model has nothing to say about such types of behaviour Agents are active –A multi-agent system is inherently multi-threaded, in that each agent is assumed to have at least one thread of active control

Summary An agent is a computer system that is capable of independent action on behalf of its user or owner A multi-agent system is one that consists of a number of agents which interact with one another In order to interact successfully, agents need the ability to cooperate, coordinate and negotiate

Two key problems How do we build agents that are capable of independent, autonomous action in order to successfully carry out tasks that we delegate to them? How do we build agents that are capable of interacting (cooperating, coordinating, negotiating) with other agents in order to successfully carry out the tasks that we delegate to them, particularly when other agents can not be assumed to share the same interests/goals?