Agent Based Brokerage Service in Electronic Commerce
ABROSE aims to create an agent based brokerage service. The main objectives of the project are to:
design, implement and integrate the ABROSE system, relying on an existing agent platform, and incorporating the main features:
dynamic knowledge capture,
multi agent system for knowledge representation,
user agent support for requests and navigation,
provider agent support for offer registration and propagation and
improving the performance and the quality of mediation through usage of collaborative adaptive agents.
verify, demonstrate and evaluate, the technical and economic suitability of multi-agent technology for commercial applications,
demonstrate how agent technology helps to create customisable and adaptive applications in the information brokerage domain,
evaluate the projects results, provide recommendations of refinements and enhancements for Multi Agent Facilities, to the relevant standard fora such as OMG MAF and especially FIPA.
Figure 1: Potential usage of agents in Information Brokerage
ABROSE takes into account the results from the AC206(ABS) and AC230 (COBRA) projects, relying on
existing models for the Brokerage Service
ABROSE will include:
final design and implementation
of the agent based architecture illustrated in figure 2
international demonstrators and
an accompanying evaluation.
Figure 2: Functional blocks of the ABROSE architecture
Summary of Trial
ABROSE will demonstrate its system at two different European sites. One of these is the FT/BATRU Lannion site, the other is the Onyx site, which is TCP/IP based and publicly available. ABROSE will apply the system in two different domains:
Tourism: Degriftour, an electronic travel agency, will apply the system to offer customers a remote service and to allow providers to promote special offers to their customers. The demonstration plans to allow test users to access a travel agency brokerage service, which will provide direct brokerage.
Teleworking: Teleworking is a common way of distributing work. Relying on an existing service offered by Onyx on their site, ABROSE will demonstrate how an intelligent matching of clients' requests for work to be done and offers from brokers and teleworkers can improve the brokerage service.
Brokerage services address the needs of users and content providers, as they facilitate the conjunction between the user's demand and the content provider's offer.
Multi Agent Systems
In a Multi Agent system, each agent has a local view of the environment, has generally specific goals and is unable by itself to solve the global task allocated to the system. For most application tasks, it is extremely difficult or even impossible to correctly determine the behavioural repertoire and concrete activities of a multi-agent system a priori, that is, at the time of its design and prior to its use.
Multi Agent Systems in Information Brokerage
ABROSE proposes a structure composed of two levels, each level corresponding to a type of particular agent:
is a low-level agent. It contains and manipulates users' and services' knowledge it represents. A TA is added when a new service is created or when a user makes his first connection. A TA has beliefs in the skills of other TAs which are on the same site. They communicate by exchanging information on the field. When a TA receives a message, it interprets it. They have a co-operative social attitude defined by the designer of ABROSE.
is a high-level sedentary agent. It contains numerous TAs. A MA is created when a new site is installed. They control creations, suppressions and migrations of the TAs which compose it. Their skills are the skills of their TAs and the beliefs they have in their TAs. As a TA, an MA has a local view and has no direct beliefs in TAs which are in others MAs. They have a corpus of beliefs about other MAs which is initially fixed in order to guarantee the convexity of multi-agent system ABROSE.
The expected achievements of ABROSE are:
A full operational prototype,
API delivered as a class of objects
enabling application programmers to define and deploy Brokerage Services,
Contributions to the FIPA standardisation body,
An overall techno-economic evaluation of the technology.
to demonstrate how agent technology can be used to cope with the complexity of the future information infrastructure,
to demonstrate how agent technology can help in dealing with the complexity of network use, as it enables more intelligent use of technology as well as better adapted brokerage services and applications,
to evaluate the project's results, provide recommendations of refinements and enhancements for Multi Agent Facilities, to the relevant standard fora , mainly FIPA.