Extension processing in a parallel and distributed environment

Extension set theory is developed for deriving solutions and transformations in problems, especially for contradictory problems, in an intelligent system. A matter element (ME) which contains three tuples, the name N, the characteristic c and the value ν, is defined to represent an object in the extension set theory. By means of continuous substitution and estimating the correlation between MEs found in the database, a problem is solved or transformed to an equivalent problem which is easier understood and solved. Due to the huge computing power consumed in the process of transforming and comparing MEs and estimating the correlation between the MEs found. Powerful and more effective schemes and architectures are required. We propose a parallel and distributed environment for extension processing. The kernel of the system is an architecture based on dataflow execution with a technique of parallel prefix computation, which is the model of a predefined macro actor to perform the extension process. The extension processing or continuous performing is spawned simultaneously when the associated MEs are returned from the database: the returned MEs are processed in a parallel and distributed manner. Therefore, a higher performance and throughput can be achieved. The evaluation of the scheme is developed and tasted according to the requirement of the extension process. From the simulation results, the proposed scheme and associated architecture can provide an effective environment for transformation of MEs