Agent-based Applications and Technologies

Asko Komsi
Nokia Research Center

Application 1: Conflict Management

The purpose of these agents is to solve conflicts based on the information the agents can collect and evaluate. Examples of such applications are:

- Dynamic scheduling of calendars, conferences, applications etc.

- Management of multiple (virtual private) networks

- Management of QoS, bandwidth allocation etc. for network elements

Application: Collection of Information and Filtering

The purpose of information filtering agents is to take any information that is provided (automatically given or collected by the agent) and filter it to provide only the minimal amount of information that is of interest to the user or the application. Examples of such applications are:

- Mail filters

- Web page preloaders

- Proxies for filtering (compressing/decompressing, lowering the picture resolution etc.) Web content

- Search robots

- Home shopping and the search for goods

Application: Avatars

The purpose of avatars is to act as reincarnations of the users while they themselves are for one reason or another not available. Examples of such applications are:

- Representation of the mobile user while roaming

- Automatic recording of preferred TV programs etc.

- Automatic control of home appliances

Agent Technologies

These are the basic technologies that are needed for the previously mentioned agent applications:

- Portable (different platforms, different memory requirements, different CPU speeds, different communication capabilities etc.) agent platform where agents can be executed and where agents can relocate from one place to another if there is a need for that. Part of this is the specification of an agent framework (the wrapper for agent state and internal logic with the interfaces to the platform) for representing the agents, too.

- Agent-to-agent communication protocols. The agents should be able to communicate with other agents. A common protocol or a set of common protocols should be available for initiating such communication and transferring information.

- Vocabularies for different agent-to-agent communication. The protocol to transfer information is not enough for meaningful communication. The agents should be able to represent and exchange information of all possible areas where agent technology is viable. This means dedicated vocabularies for each specific area of information.

- Security: authentication, verification and encryption for making sure that the agents are what they claim to be, they have not been tampered and all agent related information can be kept secret if so required. Based on these features different information access schemes can be implemented.

- The initial set of agents for helping the other agents to fulfill their tasks. These agents could act as information brokers, helpers, links to other systems etc.