Lost Wax Agent Framework

Multi-agent systems require a host container to allow agents to be created, execute and interact with other agents. These containers provide services such as creating and naming agents, discovery of other agents, support for asynchronous messaging between agents and management of the agent lifecycle.

The Lost Wax Agent Framework provides a powerful and easy to use environment for the design, development and deployment of multi-agent systems.

Current systems have been mainly originated from academia reflecting the background of research for which they were developed. Lost Wax have created one of the first systems specifically designed for the commercial enterprise. The Lost Wax Agent Framework provides a powerful and easy to use environment for the design, development and deployment of multi-agent systems. It has been designed from the start for scalability and allows the parallel execution of hundreds of agent processes distributed across multiple servers.

Logical Architecture

The framework supports integration with external systems, whether other agent frameworks or conventional systems. It supports both persistance and transactional capability. It also includes powerful monitoring and logging tools to support the run time support of deployed systems.

Agents are often used in particular domains such as logistics and supply chain. The Framework offers specific support for these areas by providing template libraries and providing functionality for creation of visual views of the applications.

Industry Standard Technologies
The Lost Wax Agent Framework is built on top of industry standard technologies, including J2EE and RDBMS. It provide an environment for agent applications to execute, communicate and interact with outside systems.

Intregration with Agent Business Analysis
Lostwax has developed a methodology for agent analysis and design based on existing agent methodologies and adapted for the development of commercial applications.

The Lost Wax Agent Framework supports this methodology by the use of case tools to visually define the roles and interactions which are then automatically converted to classes in the application. Agent libraries are provided defining standard agent protocols such as contract net, auctions and negotiations. The framework also includes a number of architectual templates which can be included as reasoning models for an agent such as state based, BDI and rule based models.