Development of Parallel Multiobjective Genetic algorithm for Designing Flexible Systems
| Project/Area Number |
17500100
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Doshisha University |
|---|
Principal Investigator |
HIROYASU Tomoyuki Doshisha University, Faculty of Engineering, Associate Professor (20298144)
|
|---|
| Project Period (FY) |
2005 – 2007
|
| Project Status |
Completed (Fiscal Year 2007)
|
| Budget Amount *help |
¥3,810,000 (Direct Cost: ¥3,600,000、Indirect Cost: ¥210,000)
Fiscal Year 2007: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2006: ¥700,000 (Direct Cost: ¥700,000)
Fiscal Year 2005: ¥2,200,000 (Direct Cost: ¥2,200,000)
|
|---|
| Keywords | Flexible System / Multi Obiective Optimization / Genetic Algorithm / Parallel Processing / PC Cluster / Grid / 多目的 / グリッド |
|---|
| Research Abstract |
In flexible system, values of target object can be changed with small number and little change of design variables. In this research, the concept of the flexibility of design variables to Pareto-optimal solutions in Multi-Objective Optimization problems was proposed. The following procedure is a method for evaluation of the flexibility of design variables to the Pareto-optimal solutions.
1) Formulate the multi-objective problem. This problem is solved using all the design variables. In this case, no design variable values are fixed.
2) The derived Pareto-optimal solutions are sorted along with the value of a certain object function.
3) Five points are determined from the sorted Pareto optimal solutions. The distances between the five points should be the same.
4) In each point, one of the design variables is changed. In this case, the values of the other solutions are fixed. The range of the design variable is assumed. In this condition, the multi-objective optimization is performed again and the new Pareto-optimal solutions are derived. These new Pareto-optimal solutions determined along with the changes in the parameter are equal to the flexibilities of this point.
5) The flexibilities of all design variables and all five points are derived.
From this procedure, the flexibility of the design variable to the Pareto-optimal solution is derived. The proposed method was applied to diesel engine fuel emission scheduling problems and it is confirmed that the Pareto-optimal solutions can be derived with only the design variables whose flexibilities are high.
|
Report
(4 results)
Research Products
(71 results)
