壅塞折扣之適用性分析

A Feasibility Analysis of Congestion Discount

林士凱(Shyh-Kae Lin)

指導教授 ： 呂育道

國立臺灣大學/電機資訊學院/資訊工程學研究所/碩士(2009年)

https://doi.org/10.6342/NTU.2009.01425

摘要

電信網路服務中常有尖峰時段產能供給不足，離峰時段卻供過於求的問題，壅塞折扣(congestion discount)是藉由提供折扣為誘因，將尖峰時段使用需求移轉至離峰時段以處理這類壅塞控制(congestion control)問題的方法之ㄧ，本研究將探討壅塞折扣的適用性及定價問題。
不同的電信服務或不同的使用時段都可能有不同的網路使用形態，其間的差別可能在於尖峰、離峰的使用量差異或者平均服務時間的長短，本研究嘗試利用等候理論中M/M/c的模擬方式了解壅塞折扣的遞延機制適用哪一類的使用型態，以及了解哪一類的遞延規則可以獲得較好的系統使用效能，藉由操作尖峰時段流量、使用者平均服務時間及折扣接受率等變數，觀察總服務時間、服務阻斷情形及服務遞延情形等結果，了解此遞延機制的適用性。過程中引進直交表(orthogonal array)的實驗設計方法，在不失準確性且保留變數間交互作用的前提下，提升實驗的效率。
關於訂價的部份，由於使用者可能透過長時間累積的使用經驗，預測連線時可能獲得此折扣的尖峰時段，而於該時段連線只為獲取適用下一時段的折扣，本研究藉助賽局理論中「純策略均衡」及「混合策略均衡」兩種方式，解釋如何調整折扣給予機制以降低這類問題的影響。
最後，經由上述的分析，本研究得到下列結論：
一、適用性分析
(1)壅塞折扣配合適當的遞延規則可適用不同的網路使用形態。
(2)相同遞延規則若同時適用尖峰及離峰時段的使用者，可能造成系統利用率降低的反效果。
二、折扣定價
(1)藉助賽局理論的混合策略均衡，可以求得網路服務供應商於壅塞時段正常收費及給予折扣的比率分配、使用者當期上網需求和只為獲取下期折扣之間的比率分配以及折扣定價方式。
(2)透過變更遊戲規則的方式，可降低使用者於尖峰時段連線只為獲取下期折扣的動機。
三、混合機制：利用門檻較高的遞延規則，於系統即將滿載時，提出折扣交易的請求，搭配能夠達到較高折扣接受率的折扣定價策略，便能提昇系統效能利用率。
整體而言，針對不同的網路使用形態，配合適當的遞延規則，壅塞折扣能提升網路的使用效率及網路服務供應商的效益，尤其是減緩因網路壅塞對客戶滿意度的影響。

關鍵字

壅塞折扣 ； 壅塞控制 ； 等候理論 ； 定價 ； 賽局理論 ； 混合策略均衡

並列摘要

Congestion discount is one solution for congestion control. The scheme is intended to shift demand during congestion periods to periods of lower demand by offering price discounts as an incentive to users to delay service during high demand periods. The price-discount offer acts as a congestion-avoidance scheme that also balances communication traffic across different time periods. This research focuses on the feasibility study and pricing problem of this scheme.
Different network environments have different features, such as different average holding time, different flow gap between peak and off-peak. These features may impact the feasibility of congestion discount. Therefore, we develop an M/M/c model in queuing theory to simulate this scheme for this feasibility study. This simulation uses flow of peak period, average holding time, and price-discount acceptance ratio as input. The output of this experiment includes total service time, the proportion of requests blocked, and proportion of requests deferred. The popular method of experimental design, orthogonal array, is used in this research in order to take interaction of different input into consideration.
User of this network service is possible to forecast the peak period, in which he can get discount for next period. Therefore they may connect the network for this discount, not for network service in current period. In the long term, this behavior will make the service provider lose revenue. In this thesis, we propose a pricing method of congestion discount with mixed strategy equilibrium in game theory to eliminate the incentive of this behavior.
Finally, the main findings of this thesis are as follows.
1.Feasibility study: The scheme with suitable deferring criterion is workable for different network environments. If the same deferring criterion is applied to peak and off-peak in one network, the utilization of the network could be declined.
2.Discount pricing: mixed strategy equilibrium and changing the rule of offering discount are both useful to eliminate the incentive for getting discount, not for network service.
3.Mixed mechanism: The utilization of network can be improved by a strict deferring criteria and a discount with high acceptance ratio offered when the congestion is imminent.
The congestion discount with suitable deferring criteria not only can improve the utilization of network and benefit of service provider, but also can mitigate the impact to customer satisfaction during network congestion.

並列關鍵字

congestion discount ； congestion control ； queuing theory ； pricing ； game theory ； mixed strategy equilibrium

參考文獻

2.Eitan Altman, Dhiman Barman, Rachid El Azouzi, David Ros, & Bruno Tuffin, “Pricing Differentiated Services: A Game-Theoretic Approach.” Computer Networks, 50, No. 7 (2006), p 982－1002.

3.Matthias Falkner, Michael Devetsikiotis, & Ioannis Lambadaris, “An Overview of Pricing Concepts for Broadband IP Networks.” IEEE Communications Surveys , Second Quarter 2000, (2000), p 12.

4.Neil Keon & G. Anand Anandalingam, “A New Pricing Model for Competitive Telecommunications Services Using Congestion Discounts.” Informs Journal on Computing, 17, No. 2 (January 2005), p 248－262.

5.Roger A. Mccain & 陳建良 譯, Game Theory：a Non-Technical Introduction to the Analysis of Strategy, Thomson Learning. 2006.

6.Garg Rahul, Kamra Abhinav, & Khurana Varun, “A Game-Theoretic Approach Towards Congestion Control in Communication Networks.” ACM SIGCOMM Computer Communications Review, 32, No. 3 (July 2002), p 47－61.