National Institute of Technology, Karnataka

Major Project Report

Dynamic Pricing of call rates

Submitted By

Amit Rajawat Anuj Jain Chetan Dugar

Submitted to

10CO11 10CO32 10CO106

Dr. Swapan Bhattacharya

Department of Computer Science and Engineering

Department of Computer Science and Engineering National Institute of Technology Karnataka, Surathkal Srinivasnagar, India 575025

Contents
1 An overview of Dynamic pricing for call rates 1.1 What is Dynamic Pricing? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 An overview of Prior Work 3 Scope of our work 4 5 Problem Statement Problem Addressing 5.1 Game Theory Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Challenges faced . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2 3 5 6 7 7 8 9

6 Evaluation Matrix

An overview of Dynamic pricing for call rates

In the current scenario we come across at-rates for calls made using any cellular network. This is clearly neither the best choice for the user nor for the network provider. The network provider wants to make the most ecient use of the bandwidth and make the maximum prot out of it. So, he wants more users to use his network when it is under-utilized. Similarly at the time of congestion, he knows that the demand for his service has increased but he doesnt have any means to make extra prot out of it. Similarly, the user wants to have the best possible service at the minimum possible price. But in static pricing mechanism there is no incentive given to user to make calls when the network is underutilized. And at the peak hours of the network, he faces a degradation of the quality of service in-spite of paying at the same call rate. So, we propose a solution to this problem using the mechanism of dynamic pricing. This is protable both to the user as well as the service provider. We show the use of the dynamic pricing at the extreme ends of the congestion in the network (over-congested and under-utilized). Prot of this mechanism is obviously witnessed even in the average congestion hours. At the peak hours of congestion, users face a degradation of QoS. The network providers even have to block calls to end the congestion. Using dynamic pricing, the users shall be guaranteed asked level of service and the providers will be benetted with larger revenue due to increase in call rates coming from higher demand. When the network is under-utilized, the bandwidth of the network is wasted. So, the reduce in call rates due to the reduce in demand will increase the willingness of the users to pay. Thus bandwidth shall be eciently utilized and network provider shall earn larger prot.

1.1

What is Dynamic Pricing?

Dynamic pricing is an approach which facilitates network provider to regulate the network usage by controlling the Quality of Service to the exising users on network and varying the call rates to the new calls to ensure optimal usage of network resources. From subscriber point of view it provides options of switching among dierent networks available which best matches with users needs thus avoiding monopoly rates of any single network provider. Dynamic pricing strategy Allows providers to recover their operating costs and nance future capacity expansions It can lead to a more ecient use of the network resources by providing sucient incentives to users, and Moreover, enables the creation of a healthy market environment, where new network services can be introduced and sustained. Several authors have applied economical concepts in the development of algorithms for mobile and wireless networks . Service providers consider factors like simplicity, scalability and the feasibility of implementation, in the selection of a pricing schema . On the other hand, users demand pricing and charging schemes more adaptable and accurate. In this way, the users behavior can be regulated and the network management is signicantly improved.

An overview of Prior Work

Lot of work has been done on dynamic pricing of call rates taking dierent scenarios into consideration. A major part of the research in [1],[4],[5] has been to create a pricing scheme taking only one network provider into consideration. The paper [5] by Dat Vo and Matthew Sorell presents a good approach on network provider which oers subscribers a continuous non-interrupting service by altering QoS of the call depending upon the network congestion. It states the idea of Willing to Pay(WTP) and Willing to Bid(WTB) method to improve current QoS. If the subscriber is willing to pay higher, the QoS would be increased to 1 quantum or if he can bid to particular value then according to that the best QoS will be provided. When a subscriber is admitted to the network, a timer for the guaranteed period for the admission QoS and price is set. When this timer expires, if the overall load of the cell exceeds allowed load threshold (Lcit), degradation in QoS of the call will take place. This applies to all active subscribers whose guaranteed QoS period has expired. The reduced QoS level will be calculated based on three factors: the number of active subscribers who are still in the guaranteed period, the thresholdLcit, and the number of active subscribers whose guaranteedQoSperiod has expired. Active subscribers whose guaranteed period has expired can bid up for an increase in QoS for their call. This behavior reects the characteristics of thecurrent user group in the cell. Reference [4] proposes scheme which employs dynamic pricing during all network conditions. It works as follows. The CAC component continuously monitors the amount of available bandwidth (i.e., unutilized bandwidth). When the amount of available bandwidth changes due to call completion or new admitted calls, the CAC component then computes the optimal arrival rate for each QoS class in order to maximize the utilization of the new available bandwidth in the system and achieve certain fairness levels between QoS classes. The actual arrival rates for the QoS classes are, however, dierent from the optimal rates determined by the CAC component. In this case, the dynamic pricing component dynamically determines the prices of units of bandwidth for each class based on the users demands in order to force the actual arrival rates to be less than or equal to the optimal ones. The dynamic prices are computed independently from the optimal arrival rates, hence, simplifying the implementation of our scheme and providing the network operators the exibility to use any CAC and users demands functions without aecting the computation of prices.Described scheme is able to maximize the utilization of Broadband Wireless Access Systems (BWASs) when these networks are underutilized as well as preventing congestion when they are over utilized. Moreover, scheme supports multiple QoS classes with calls having multiple bandwidth requirements, which makes it more suitable for BWASs. Reference [1] considers both (service providers and users satisfaction) by using a exible QoS based pricing approach for heterogeneous wireless access networks. It includes a selection mechanism that allows users to choose the wireless access network that best adapt to their needs, and preserves the QoS level during the time the service is being used. Additionally, QoSDPA algorithm denes a dynamic pricing strategy that is applied when the QoS level is degraded. This strategy attempts to maintain a high level of users satisfaction, while maximizing the prots obtained by service providers. In [6], a pricing policy for multiple competing ISPs (Internet Service Providers) using a threat strategy is presented. Ref. [7] shows a proposal in which authors use game theory to analyze the impact in the cost based on the economic interests of a wireless access point owner and his/her paying client. In Ref. [8] several problems for resource allocation and base-station assignment in CDMA (Code division multiple access) networks are studied. But our area of concern is an oligopoly market having multiple network providers serving multiple users. Reference [3] proposes the above mentioned scenario. It takes into consideration two levels of service: Voice service and Data service. They show that making the decision of pricing unchangeable for voice service after the initial setting will lead to malicious use of the network by users. Similarly keeping the option of price change at later change can lead to malicious network provider strategies. Using a non-cooperative game model, they proposed dierentiated pricing strategies for the WSPs and users. They also propose an optimal strategy for access network selection by the service providerswhich minimizes the eect on QoS of the existing users and increases self co-existence. They showed that in order to achieve Nash equilibrium, it is in the best interest of both WSPsand users to adhere to the proposed strategies. But the problem with their strategy is that they have made the decision of creating sub-games as 3

static i.e. we need to decide before itself as to how frequently the pricing decision needs to be made. There is no adaptable solution that is provided to it. Besides this, there is no mention as to how can we approximate the competitor network providers pricing decision. They have also assumed that every provider shall provide the same service and this shows that every provider shall end up putting the same price by the end in order to avoid losing the game. We would like to incorporate, dynamic way of deciding as to how long a user must be given assured service without interruption. We would also approximate the contending providers pricing strategy while taking the pricing decision.

Scope of our work

Users are willing to play a more central role by modeling and reshaping the service experiences upon their needs. As mentioned, our model is user-centric which improves the service oering protable, value-added services faster and cheaper than ever before. In this scenario we assume that end-users have more freedom connecting dynamically to any WSP for any service through their multi-mode devices and can disconnect at anytime, which means they do not have any strong relationship with service providers. So WSPs and end-user only interact on a per-service or per-session basis. From a users perspective, the objective is to select the WSP that best adapts to his/her needs. From WSP perspective the objective is to dene dynamic pricing strategies for each service and the requested QoS level; this is important to consider because if prices are low, WSP would attract too many users leading to degradation in performance due to resource sharing. At the same time, if advertised prices are too high, it will lead to users turning down that WSP. For the analysis we have considered that only voice service is provided by the network service providers. Every service provider oers the information of call rates per bandwidth, but the call rates per bandwidth would also depends on how much time the user wants the assured quality of service. The rate obviously shall be more if user bids it for larger time. The network service provider shall decide this change on call rates for dierent time period of assured service on the basis of congestion in the network. Development in wireless access technologies and sophisticated personal user devices are driving the way towards a heterogeneous wireless access network (HWAN) environment which will enable anytime and anywhere communication.This new market will promote the generation of a multitude of wireless service providers (WSPs), which can use a mix of wireless access technologies to provide these new services to end-users with competitive prices. HWAN will change the traditional vision of service model from network-centric to user-centric.Unlike traditional service model where all management functions are controlled according to the service providers perspectives, the usercentric vision for the future considers end-users in a heterogeneous wireless network environment where they will be free to shop around not only for the service they need, but also for the available wireless access network which meets their current service needs. Users will take advantage of this competition scenario and they will always connect to the network that can best satisfy their needs and preferences for the current application using novel algorithms to select the optimal access network in an intelligent way according to their needs. In this situation, heterogeneous wireless networks look for the market evolution from traditional monopolies that dictate usage conditions, to a user-centric service-oriented environment. However, within the current service pricing and provisioning model, there is still a strong relationship between end-users with a single wireless service provider. End-users get services from one provider for a period of time based on the contractual agreement. Prices for the various services are based on the charging model for voice services (i.e. free, at rate or any of their variations). In the at-rate model price are xed and do not uctuate according to changes in network conditions. The price is paid monthly for the usage of a certain facility. However, the current demands of quality of services into wireless service access have demonstrated the need for more sophisticated pricing strategies .We want to present a QoS-based dynamic pricing approach for service provisioning in a heterogeneous wireless network environment where requirements of new services demand ecient and exible pricing strategies and charging mechanisms. We want to propose an adaptable scheme to changeable environments, which satises users demands while maximizing service providers revenue. We want the users to have the exibility of even deciding the time for which the quality of service is assured.

Problem Statement

In our problem statement we have determined following input parameters to our system to calculate the output value for concerned network service provider. We assume that every network service provider would be provided with certain input parameters (number of users in each network) to precise come up with most optimal output. Input parameters: Number of network providers N, Number of existing users M, Congestion of concerned network at time t, Number of users in each network. For each user we investigate the following information: if he is using any network amount of bandwidth consumed by user time period of promised service usage call rate per bandwidth Output: Call rate per bandwidth of concerned network. It is a sequential system which uses previous output value as one of the current input value to calculate the current output. The call rates shall be decided using a game theoretic approach. We shall x on a pricing scheme which is the maximum price that our concerned service provider can impose and still end up as the winner of the game involving other service providers. The game consists of N service providers such that each has a particular strategy Si and a utility function U. The utility function takes into consideration both the Service Providers revenue along with the cost of eort or resources to get that revenue. So it is his net utility. But a very important concern with our suggested system is that in any case the user must also be having the maximum utility so that it sticks to our Service Provider.

Problem Addressing

To come up a solution to the problem we would use a non-co-operative game theoretic approach. So we would like to dene terms like utility function for the user, revenue for the network provider, user satisfaction, change in quality of service as per the equation in [3] U S (b) = 1+e 1 bmax bmin 2
b

(1)

where US(b) is the user satisfaction perceived for bandwidth b,bmin is the minimum bandwidth required to maintain the service and bmax is the maximum bandwidth above which the user perceives no signicant improvement in the QoS. Similarly, we can dene user utility, US(b)= user satisfaction price paid Service provider utility = price paid by the user. QoS aect ratio is the ratio of bandwidth allocated Ba under a network to the total bandwidth capacity B of the network denoted as from [3], Qx = Ba B (2)

Now the challenge will be to nd the sub-games involving users and network providers following non-co-operative strategy so that Nash equilibrium could be established between them. For applying game theory we referred [2]. So we suggest the following approach for our solution:

5.1

Game Theory Approach

While the Quality of Service (QoS) oered to users may be enhanced through innovative protocols and new technologies, future trends should take into account the eciency of resource allocation and network/terminal cooperation as well. Game theory techniques have widely been applied to various engineering design problems in which the action of one component has impact on (and perhaps conicts with) that of any other component. Therefore, game formulations are used, and a stable solution for the players is obtained through the concept of equilibrium. Two types of games are distinguished: in non-cooperative games, each player selects strategies without coordination with others. On the other hand, in a cooperative game, the players cooperatively try to come to an agreement, and the players have a choice to bargain with each other so that they can gain maximum benet, which is higher than what they could have obtained by playing the game without cooperation. Our concerned game is clearly a Non-Cooperative Game. Nash Equilibrium: In game theory, the nash equilibrium is a solution concept of a game involving two or more players, in which no player has anything to gain by changing only his own strategy unilaterally. If each player has chosen a strategy and no player can benet by changing his strategy while the other players keep theirs unchanged, then the current set of strategy choices and the corresponding payos constitute a nash equilibrium. So our goal is to reach to this Nash Equilibrium. Mixed Strategy: When a player makes a decision, he can use either a pure or a mixed strategy. If the actions of the player are deterministic, he is considered to use a pure strategy. If probability distributions are dened to describe the actions of the player, a mixed strategy is used. We need to work for this scheme as the oppositions strategy is not known. Repeated games: In strategic or static games, the players make their decisions simultaneously at the beginning of the game. On the contrary, the model of an extensive game denes the possible orders of the events. The players can make decisions during the game and they can react to other players decisions. Extensive games can be nite or innite. A class of extensive games is repeated games, in which a game is played numerous times and the players can observe the outcome of the previous game before attending the next repetition. Our game is divided into numerous sub games. So, repeated games is clearly required. Even if we make prot in the current move, its not assured in the next move. 7

Clearly we apply our strategy only in the Transport Layer for:Call Admission Control, Load Control, Cell Selection. But we would be concerned only with Call Admission Control and Load Control. There are two games that we care about: Provider Vs Provider which can result in just one winner and Customer Vs Provider in which we are more concerned in the comparison of utility achieved by the user when compared to unavailability of our proposed system. The proposed game is non-cooperative because, on the one hand, the service providers wish to maximize their revenue and, on the other hand, the users wish to maximize the quality of service received, keeping at the same time the expenses as low as possible. Since these two goals are obviously contradictory, the players do not have the slightest motivation to cooperate. The game is also nonzero-sum, since an increase in one players payo does not imply a decrease in the other players payo.

5.2

Challenges faced

Assumption of rationality Game theory is founded on the hypothesis that each player plays rationally and thus seeks his best interest in a rational manner. When dealing with nodes or terminals however this behavior cannot be always guaranteed. Assumption of willingness to cooperate In cooperative games it is assumed that players will collaborate in order to maximize their prots. A signicant problem is that players sometimes choose to behave selshly or even cheat in order to optimize their own prot. For this reason, in certain occasions, incentive mechanisms for cooperation, as well as disincentives against cheating need to be formulated. Choice of utility functions/ payo calculation This is unquestionably the most challenging part of a game-theoretic framework, since the utility function interprets the players perception of performance and satisfaction. Utility functions also show the trade-os the player is willing to make, usually between acquiring more resources and saving money. Not guaranteed existence of equilibrium In game-theoretic formulations an analysis is often required to check if they reach a nash equilibrium. Even if an equilibrium is reached however, the existence of multiple equilibria is not always excluded. In such case the most ecient and stable one has to be sought. Cooperation Incentives: In several cases, optimization issues require collaboration among nodes. Nevertheless, cooperation cannot be taken for granted; even though in most cases players do obtain the optimal result by sharing resources with others, in certain cases it is not clear enough for them why they should not act selshly or even not try to cheat. Nodes exhibiting such behavior are termed selsh and malicious correspondingly. The basic idea for node punishment is that nodes should be rewarded or penalized based on their behavior. Nodes that oer resources should be aided. On the other hand, selsh nodes should be gradually isolated from the network.

Evaluation Matrix

In this section we compare dynamic pricing versus static pricing from the perspective of utility obtained by the users and the network providers. Then the improvement in user satisfaction due to guaranteed time period for bandwidth demanded. Our approach species three policy classes: access, negotiation, and pricing and charging. The main objective of the policies is to satisfy users service demands by rstly, maximizing the provided QoS level; and secondly, when the QoS level cannot be provided, by applying policies with fairer taris attempting to recover the users satisfaction level as soon as possible. These measures should transcend in a set of benets for the service providers. Utility Function: The category of service parameters refers to information of service requirements; the category of network parameters is related to the current network conditions; and nally, the last category of users preferences represents relevant values for some factors as price and QoS level. Users may specify the importance or weights of each parameter which sum to 1, i.e., if the price is more important than the level of quality of service, then the user might dene the values for price and QoS level as 60 % and 40 % respectively. Service parameters used in the utility function is mostly bandwidth. But we would also like to incorporate other parameters like congestion, delay, jitter, error rate etc. The system is clearly capable of including other such service parameters. Our entire result shall be based on the utility obtained through our proposed scheme. The net utility obtained by the users must be maximum. Even the concerned provider must be the most protable provider amongst the concerned competitors. For showing that the user has maximum utility, we would try to come up with a formidable proof when its compared to static pricing scheme. For the multiple network environment, we would like to show through simulation that the concerned service provider always wins. If we come up with a optimum solution then we can even guarantee our result even without any simulations.

References
[1] A QoS-based dynamic pricing approach for services provisioning in heterogeneous wireless access networks by Antonio Guerrero-Ibez, Juan Contreras-Castillo, Antoni Barba , Anglica Reyes Digital Object Identier: 10.1109/IE.2010.49 Publication Year: 2010 Page(s): 231 - 236 [2] A survey on game theory applications in wireless networks by Dimitris E. Charilas, Athanasios D. Panagopoulos Computer Networks Volume 54, Issue 18, 20 December 2010, Pages 3421 - 3430 [3] Dynamic pricing for service provisioning and network selection in by heterogeneous networks by S. Sengupta, S. Ananda, M. Chatterjee , R. Chandramoulia Physical Communication Volume 2, Issues 12, MarchJune 2009, Pages 138 - 150 [4] Ecient Bandwidth Management in Broadband Wireless Access Systems Using CAC-based Dynamic Pricing by Bader Al-Manthari, Nidal Nasser, Najah Abu Ali, Hossam Hassanein Local Computer Networks, 2008. LCN 2008. 33rd IEEE Conference on Digital Object Identier: 10.1109/LCN.2008.4664208 Publication Year: 2008 , Page(s): 484 - 491 [5] Signaling Requirements for Smart Dynamic Pricing in Cellular Networks by Dat Vo and Matthew Sorell Communications and Networking in China, 2006. ChinaCom 06. First International Conference on Digital Object Identier: 10.1109/CHINACOM.2006.344750 Publication Year: 2006 , Page(s): 1 - 5 [6] Economics of network pricing with multiple ISPs by S. Shakkottai, R. Srikant Proc. of IEEE INFOCOM, vol. 1, 2005, pp. 184-194. [7] J. Musacchio, J. Walrand, WiFi access point pricing as a dynamic game, IEEE/ACM Transactions on Networking 14 (2) (2006) 289-301 [8] S.C.M. Lee, J.W.J. Jiang, J.C.S. Lui, C. Dah-Ming, Interplay of ISPs: distributed resource allocation and revenue maximization, in: Proc. 26th IEEE International Conference on Distributed Computing Systems, vol. 19 issue 2, 2006, pp. 204 -218

10