2002 Fiscal Year Final Research Report Summary
Efficient Algorithms for Large Optimization Problems
| Project/Area Number |
12680433
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
社会システム工学
|
|---|
| Research Institution | Tokyo Institute of Technology |
|---|
Principal Investigator |
MIZUNO Shinji Tokyo Institute of Technology, Graduate School of Decision Science and Technology, Professor, 大学院・社会理工学研究科, 教授 (90174036)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
NAKATA Kazuhide Tokyo Institute of Technology, Graduate School of Decision Science and Technology, Assistant Professor, 大学院・社会理工学研究科, 助手 (00312984)
YAJIMA Yasutoshi Tokyo Institute of Technology, Graduate School of Decision Science and Technology, Associate Professor, 大学院・社会理工学研究科, 助教授 (80231645)
|
|---|
| Project Period (FY) |
2000 – 2002
|
| Keywords | Optimization / Algorithm / Interior Point Method / Linear Programming |
|---|
| Research Abstract |
This research aims to develop efficient algorithms for solving large optimization problems and to investigate their theoretical properties and practical usage. In 2000, we did basic research for the purpose. There are 3 types of variables, free variables, nonnegative variables and bounded variables. in large optimization problems. We investigated efficient algorithms which can handle all these types of variables simultaneously. In 2001, we studied algorithms for solving large problems with a special structure. Especially, we proposed a new algorithm for solving multistage stochastic programming problems. These problems could be very large when the number of stages is not small. We use the special structure of the problems and developed an efficient algorithm for them. In 2002, we investigated a new algorithm for solving large linear and nonlinear programming problems. The algorithm uses a logarithmic transformation to the variables in the problem. This transformation makes the path of centers smooth and the algorithm efficient. From some preliminary computational experiment, we found that the algorithm can solve large problems very quickly.
|
Research Products
(13 results)
