Randomized Local Policies for Scheduling Games with Multi-jobs Players
碩士 === 國立交通大學 === 資訊管理研究所 === 104 === Most research on scheduling games assumes a single-job model, where each job can be seen as a distinct player. Each player decides to assign her job to a machine to minimize her own completion time according to a local policy for ordering jobs on a machine....
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2016
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/82478591453504218586 |
| Summary: | 碩士 === 國立交通大學 === 資訊管理研究所 === 104 === Most research on scheduling games assumes a single-job model, where each job can be seen as a distinct player. Each player decides to assign her job to a machine to minimize her own completion time according to a local policy for ordering jobs on a machine. The price of anarchy as a measure of the overall performance is defined as the ratio between the social cost of the worst equilibrium and the optimal social cost. A common social cost is the sum of all weighted completion times.
In this thesis, we study multi-job scheduling games, where each player owns several jobs and can move any subset of her jobs arbitrarily to minimize the sum of weighted completion times of her jobs. We analyze weak equilibria for multi-job scheduling games with a randomized local policy and give a price-of-anarchy upper bound less than 4.
|
|---|