Abstract In place of convensional hierarchial distributed computer control systems (abr. DCCS), a newly developed uni-level homogeneous DCCS is proposed. The system is configurated with mini-computer or micro-processor based node computers with large scale main-memory (IC memory) but no disk memory (except terminal node) and a horizontal distribution bus. Node computers are organized in one-level along with a horizontal bus. From functional standpoint, each of node computer is designed to perform autonomously. This functional independance is also discussed, focussing on both functions of CRT man-machine and of material tracking. The optimum system design problems on the proposed homogeneous system are defined and formulated, with respect to task assignment and data file allocation. The actual optimization algorithms are developed on each of the problem, based on dynamic programming method and branch and bound method. The illustrative example from actual application fields and its numerical results are also presented to show feasibility and efficiency of the proposed design algorithms.
[1]
Robert Marcus.
An application of artificial intelligence to operations research
,
1984,
CACM.
[2]
Peter P. Chen,et al.
Optimal Design of Distributed Information Systems
,
1980,
IEEE Transactions on Computers.
[3]
C. Imamichi,et al.
Development and Quantitative Evaluation of Distributed Sensor Base Management System
,
1982
.
[4]
C. Imamichi,et al.
Large Scale Control System for the Most Advanced Hot Strip Mill
,
1983
.
[5]
Hironori Kasahara,et al.
Practical Multiprocessor Scheduling Algorithms for Efficient Parallel Processing
,
1984,
IEEE Transactions on Computers.
[6]
Masahiro Tsuchiya,et al.
A Task Allocation Model for Distributed Computing Systems
,
1982,
IEEE Transactions on Computers.
[7]
Thomas L. Morin,et al.
COMPUTATIONAL ADVANCES IN DYNAMIC PROGRAMMING
,
1978
.