Bidding strategies for dynamic spectrum allocation

Abstract

Dynamic Spectrum Allocation is the process of assigning spectrum licenses in terms of the chunks of the spectrum band to the Wireless Service Providers (WSPs). This allocation assignment is being done as per the WSPs' requirement and this in turn would depend on the end users' applications demand [16]. For Dynamic Spectrum Access, economic framework is needed to make the system feasible under economic terms. This process uses some kind of service pricing mechanism that the service provider can use for the acquisition of the spectrum band and requirement of the end users. In the Dynamic Spectrum Allocation scenario, the problem is to find the appropriate bidding strategies. The approach to this problem is to simulate the different cases with varying parameters. Here, the interaction between the spectrum owner and providers is modeled through auction model which has been studied in this work whereas the interaction between the providers and end users is based on the demand. This thesis presents the bidding strategies and appropriate prediction method that maximize the revenue and the profit of both the providers as well as the end users. The auction method and different bidding strategies adopted, gives the winning criteria for the providers on how many number of units to bid and the prices for these units. Prediction method for the price uses the concept of probability of winning the particular unit. Simulation results show the comparison between prediction and actual values, revenue and profits of the providers. Among the different auction and bidding methods, the Vickrey auction has been used in this work. In the Vickrey auction method, the allocation of the resources is done efficiently as compared to other methods and does provide dominant bidding strategy.