APPMPS: An Asynchronous Parallel Plastic Multiprocessing System

The idea of loosely coupled machines with a common purpose has become a staple in many academic and research institutions. In terms of parallel computing, it presents a viable alternative to classic approaches. We consider a model of a parallel computing environment using the ideas from loosely coupled systems such as those found in clusters, swarms, or neural systems. We primarily describe in this paper the nature of the processing and the organization of memory in an asynchronous parallel plastic multiprocessing system. The ideas presented are based on a mathematical formulation of sequences of symbols representing sensor output originating from some phenomena. If multiple sequences are processed, then signal fusion issues arise, which we consider. We also describe how data are processed and results are reported. We use simple examples to illustrate these ideas.