A framework for problem solving in a distributed processing environment

Abstract : The concept of distributed problem solving, or the cooperative solution of problems by a decentralized and loosely-coupled collection of knowledge-sources that operates in a distributed processor architecture is presented. Such architectures offer high-speed, reliable computation at low cost and are an effective way to utilize the new LSI processors and the developments of the recent synthesis of computer and communications technology. A conceptual framework called the contract net framework that specifies communications, control, and knowledge organization has been developed. Task distribution is viewed as an interactive process, a discussion carried on between a node with a task to be executed and a group of nodes that may be able to execute the task. This is the origin of the contract metaphor for control, where task distribution corresponds to contract negotiation. The types of knowledge used in such a problem solver are discussed, together with the ways that the knowledge is indexed within an individual node and distributed among the collection of nodes. The use of two primary types of knowledge (referred to as task-centered and knowledge-source centered) is shown. We illustrate the kinds of information that must be passed between nodes in the distributed processor in order to carry out task and data distribution. We suggest that a common internode language is required and that the task-specific expertise required by a processsor node can be obtained by internode transfer of procedures and data.