A study on developing a distributed problem solving system
Eun-Gyung Kim · 2002
I propose a distributed problem solving (DPS) model where a complex artificial intelligence problem is solved by multiple distributed agents, each agent solves a part of a larger problem in parallel and cooperates and communicates through 'virtual shared memory(VSM)'. Also, in order to develop the DPS systems based on the proposed model, I have designed a DPS system development environment called 'DPS-Linda', using Network Linda which supports communication and cooperation through the VSM by providing a single common VSM to all of the agents within a DPS system. In DPS-Linda, each agent consists of a separate, individual interpreter (inference engine), knowledge base, and VSM manager. Each agent's knowledge is represented in the form of OPS5-like syntax of production system. I compare the features of DPS-Linda with some other DAI programming shells.