2016 Fiscal Year Research-status Report
Efficient Evolutionary Algorithm of Multi-objective Optimization for High-Confidence Cyber-Physical Systems
| Project/Area Number |
15K00120
|
|---|
| Research Institution | Japan Advanced Institute of Science and Technology |
|---|
Principal Investigator |
リム 勇仁 北陸先端科学技術大学院大学, 先端科学技術研究科, 准教授 (90435793)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
丹 康雄 北陸先端科学技術大学院大学, 先端科学技術研究科, 教授 (90251967)
|
|---|
| Project Period (FY) |
2015-04-01 – 2018-03-31
|
| Keywords | Cyber-Physical Systems / Real-time System / Task Dependency Method / Safe-to-process / Concurrent Arrival Issue / Time Delay Model |
|---|
| Outline of Annual Research Achievements |
In fiscal year 2016, the aims are to derive a mathematical representation of the time delay model with consideration of network time delay, latencies of sensors and actuators, and clock error for the HiCoCPS system. The comparison analysis of the proposed time delay model with the safe-to-process algorithm is testified through computer simulation. The concurrent arrival task issue is formalized as a critical problem of FIFO approach when those two tasks have a dependency of each other. Therefore, a task dependency method is proposed to solve this problem in the HiCoCPS system.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
The time delay model by taking advantage of task dependency method is proposed. The numerical evaluations with hybrid systems, which consist of a main controller and sub-controllers are further examined according to the planned fiscal year in 2016. Because the numerical performance of the proposed time delay model is conducted using simulation only, the cross-checking in between simulation and emulation with different test environment is not conducted. Furthermore, emulation studies under a specific application environment like smart homes are not examined. Therefore, the scalability and feasibility of the proposed time delay model are not justified.
|
| Strategy for Future Research Activity |
In next fiscal year, the scalability and feasibility of the proposed SMT-based scheduling algorithm and task dependency method are given priority to be investigated. At the meantime, the evolutionary real-time multiple objective optimization (MOO) is studied to pursue two dynamic objective functions: delay time minimization on one side, and number of completed tasks maximization on the other. And then the accomplished milestone is to evaluate and analyze the real-time operations of city energy management system of a large-scale distributed smart home environment with respect to the power distribution and generation. In the evaluation scenario, 1,000 smart homes with 4LDK type construction and the average size of each house is 120 m2. Each house consists of four actuators, i.e., air-conditioner, window, fan, and heater. In this task, the performability of the optimization model is intensively investigated.
|
| Causes of Carryover |
The leftover was occurred intentionally in order to allocate some travel expenses for the next fiscal year.
|
| Expenditure Plan for Carryover Budget |
The leftover will be used for travel expense in next fiscal year.
|
