Kanada, Y., Technical Report of IEICE, COMP92-93 and SS92-40, The Institute of Electronics, Information and Communication Engineers, pp. 1-10, 1993, Published by IEICE.
[ 日本語のページ ]
[ Paper PDF file (in Japanese) ] [ Paper postscript file (in Japanese) ]
[ OHP postscript file: Slides, Handout ]
[ OHP PDF file: Slides, Handout ]
Abstract: Systems, which perform complex symbol processing, can neigher be understood by understanding the program statically nor by tracking their microscopic behavior. Thus, macroscopic models of the computation processes are necessary to understand the systems. Macroscopic models are also necessary for automatic control of computation and realizing self-organizing computation. This paper examines foundation of macroscopic theory of computation processes, and proposes modeling based on stochastic process theory. This paper also gives an example of a macroscopic model, the Markov chain model of the Chemical Casting Model (CCM), which is a microscopic model, and analizes computation processes of a graph coloring problem based on CCM. This theory realizes a fusion of symbol computation in the microscopic model and pattern computation in the macroscopic model.
Introduction to this research theme: CCM: Chemical-Computation Model