粒子群演算法應用企業伺服器負載平衡之省電優化

Particle Swarm Optimization Algorithm Applied to Enterprise Server Load Balancing Power Saving Optimization

陳仕倫(Shih-lun Chen)

指導教授 ： 賀嘉律

國立中央大學/資訊電機學院/通訊工程研究所碩士在職專班/碩士(2013年)

摘要

在能源日漸不足及的今天，省電一直是重要議題。當今的伺服器系統可藉由智慧型電源管理技術，將整體系統作降頻動作，使資訊管理人員可以控制最佳化電源管理，使現有伺服器實現更高的價值和效能。資訊管理人員可藉由動態限定耗電量，以避免系統的過度冷卻、縮減停機時間。但由於沒完善的考量設定參數，並未正確考慮主機負載使用率，以致無法有效地提供所需要的處理運算並進行電力的節約。而粒子群演算法的特性可以求得最佳解位子，可判斷何時使用適當省電模式的級數，將可優化電力系統及效能使用率，本研究將考慮常態性一天之間，各時段之負載大小與總發電量及主機尖峰值和離峰值之匹配關係將套用於PSO演算法權重及學習因子探討方面找到最解佳負載設定值達到節能省電，以神達電腦伺服器為實例。

關鍵字

粒子群優化演算法 ； 節點管理器 ； 伺服器 ； 省電

並列摘要

In a world of energy shortage nowadays, saving energy has been on top of the issues. The servers in use can be more energy-efficient through the smart energy management system which can lower the frequency, thus enabling the IT staff to control the optimized management system in an easier way. The IT staff can dynamically limit the power consumption in order to avoid the cooling of the system and to minimize the suspension time of it. However, due to the lack of proper setup of the parameter and the lack of appropriate consideration of the load of the host, the needed operation for energy saving therefore cannot be processed. The trait of Particle Swarm Optimization Algorithm is in search of the best position, the best timings for saving energy in this case, in another word, optimizing the power system and therefore reaching better efficiency.
This study takes into consideration the load in each time period, and the maximum and minimum load of the host, adopting PSO to find the best set figure for energy-efficiency.

並列關鍵字

PSO ； Node manager ； Server ； Power saving

參考文獻

[20]洪鈺欣 - 負載預測與契約用電最佳化之研究,台灣大學工程科學及海洋工程學研究所,2011。

[1] Kirkpatrick, S. (1984), "Optimization by Simulated Annealing: Quantitative Studies", Journal of Statistical Physics, Vol. 34, pp. 975–986.

[3] E. Bonabeau, M. Dorigo, G. Theraulaz (1999), Swarm intelligence: from natural to artificial systems.

[6] Shi, Y.; Eberhart, R.C. (1999). "Empirical study of particle swarm optimization". Proceedings of the Congress on Evolutionary Computation, pp. 1945–1950.

[7] Shi, Y.; Eberhart, R.C. (1998). "Parameter selection in particle swarm optimization". Proceedings of Evolutionary Programming VII (EP98). pp. 591–600.