1992 Fiscal Year Final Research Report Summary
Applications of Optimization Techniques to Problems in Information Science
| Project/Area Number |
02452280
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
Informatics
|
|---|
| Research Institution | KYOTO UNIVERSITY |
|---|
Principal Investigator |
IBARAKI Toshihide Kyoto University, Dept. Applied Math. and Physics, Professor, 工学部, 教授 (50026192)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
MASUYAMA Shigeru Toyohashi Inst. Technology, Dept. Knowldege Information. Asso. Prof, 工学部, 助教授 (60173762)
NAGAMOCHI Hiroshi Kyoto Univ. Dept. Applied Math. & Physics. Assi., 工学部, 助手 (70202231)
OHNISHI Masamitsu Tohoku Univ., Dept. Management Science. Asso., 経済学部, 助教授 (10160566)
FUKUSHIMA Masao Kyoto Univ., Dept. Applied Math. & Physics. Assi, 工学部, 助教授 (30089114)
|
|---|
| Project Period (FY) |
1990 – 1992
|
| Keywords | Optimization / Algorithm / Distributed Systems / Graphs. Networks / Nonlinear Systems / Deductive Databases / Probabilistic Systems |
|---|
| Research Abstract |
Among the subjects of optimization encountered in information science, we have studies the following topics.
1. Optimization in graph and network theory. Based on a new linear time algorithm to decompose a given graph into disjoint forests, we developed an O(mn + n^2log n) time algorithm to compute a minimum cut of the graph.
2. Nonlinear optimization. A successive linearization algorithm is proposed for a sparse large scale nonlinear programming problem. A successive quadratic programming method is also devised for the problems that are neither differentiable nor convex.
3. Optimization of probabilistic systems. We showed that the cyclic assignment policy is optimal for a certain system of parallel probabilistic servers.
4. Optimization of distributed systems. As a method to realize mutual exclusion, we considered coteries, and established a Boolean theory to characterize various aspects of coteries. Optimal coteries in the sense of maximizing availability are also clarified for ring networks.
5. Optimization associated with deductive databases. We studied the complexity of finding optimum sequences of joins. Also an approximate simple formula is derived to estimate the complexity of computing the closure of relational tables.
|
