Stochastic Asymmetric Blotto Game Approach for Wireless Resource Allocation Strategies

Chien, S. F. and Zarakovitis, C. C. and Ni, Q. and Xiao, Pei (2019) Stochastic Asymmetric Blotto Game Approach for Wireless Resource Allocation Strategies. IEEE Transactions on Wireless Communications, 18 (12). 5511 - 5528. ISSN 1536-1276

Text (TWC-Blotto-author final)
TWC_Blotto_author_final.pdf - Accepted Version
Available under License Creative Commons Attribution-NonCommercial.
Download (708kB)

Abstract

The development of modellings and analytical tools to structurise and study the allocation of resources through noble user competitions become essential, especially considering the increased degree of heterogeneity in application and service demands that will be cornerstone in future communication systems. Stochastic asymmetric Blotto games appear promising to modelling such problems, and devising their Nash equilibrium (NE) strategies by anticipating the potential outcomes of user competitions. In this regard, this paper approaches the generic energy efficiency problem with a new stochastic asymmetric Blotto game paradigm to enable the derivation of joint optimal bandwidth and transmit power allocations by setting multiple users to compete in multiple auction-like contests for their individual resource demands. The proposed modelling innovates by abstracting the notion of fairness from centrally-imposed to distributed-competitive, where each user’s pay-off probability is expressed as quantitative bidding metric, so as, all users’ actions can be interdependent, i.e., each user attains its utility given the allocations of other users, which eliminates the chance of low-valued carriers not being claimed by any user, and, in principle, enables the full utilisation of wireless resources. We also contribute by resolving the allocation problem with low complexity using new mathematical techniques based on Charnes-Cooper transformation, which eliminate the additional coefficients and multipliers that typically appear during optimisation analysis, and derive the joint optimal strategy as a set of linear single-variable functions for each user. We prove that our strategy converges towards a unique, monotonous and scalable NE, and examine its optimality, positivity and feasibility properties in detail. Simulation comparisons with relevant studies confirm the superiority of our approach in terms of higher energy efficiency performance, fairness index and quality-of-service provision.