My research interests range over a wide spectrum of Computer Science I
am especially interested in Computational Models, Programming
Languages, Operating Systems, and Artificial Intelligence. In
particular, I am interested in Object-Oriented Computing, Concurrent
Programming, Distributed and Open Systems, and Cooperative Problem
Solving.
Recently, I have developed the following object-oriented concurrent
programming languages: Orient84/K is designed for describing
problems which are solved by multiple knowledge agents;
ConcurrentSmalltalk is based on Smalltalk-80 and is being
widely used for various applications and for teaching concurrent
programming; Paradise is based on Lisp and has been used for
describing behavioral simulation and cooperative problem solving.
Currently, my largest interest lies in the development of
computational models/paradigms for open distributed environments. As
such a model, I am proposing Computational Field Model, which
regards networks of computers as a continuous computational field. In
such a field, gravitational force, repulsive force, and
inertia/friction are taken into account for efficient computation
through grouping and load distribution. The gravitational force between
objects is defined as the frequency and quantity of communication, the
repulsive force as closeness in physical distance, and the inertia and
friction as initial and moving costs. Communication is performed in
three steps: dissimilation (or dissolution), migration, and
assimilation. Theoretical aspects of the Computational Field Model
are being intensively investigated. Also,in order to provide users
with a continuous computational field, the Muse Operating System is
being developed, where transparent object migration is provided for
efficient computation using portable and mobile computers.
In an open distributed environment, it is natural to envisage problem
solving as the cooperation of multiple agents. It is quite interesting
to consider how cooperation among agents is achieved and what
fundamental structure of each agent is. I see cooperation as performed
in the following three steps: observation, prediction through
simulation, and commitment. Cooperation can also be achieved through
conversation among agents. Among others, intention recognition and
self consciousness are then important issues. The real-time nature of
such a problem must also be considered. We are investigating
cooperative problem solving independently and in relation to
Computational Field Model to establish a new paradigm in computer
science.