Requirement of multiple services from the customers has compelled the wireless service providers (WSPs) to maintain multiple radio access networks (RANs) of different radio access technologies (RATs). Further, dynamic spectrum allocation (DSA) approach offers a great opportunity to the WSPs to utilize their available spectrum in more efficient manner. As the growth of demand for wireless services is expected to continue unabated, the WSPs may also face a scenario in which, spectrum demands from the RANs is more than the availability of spectrum with the WSP.In this work, we have addressed this limited spectrum availability scenario in the composite wireless network (CWN) considering realistic constraints such as different propagation characteristics of channels at different spectrum region and discrete channelizations of the RANs. The problem has been posed as a bankruptcy game and Shapley value solution of the game has been computed. For maintaining the fairness of spectrum allocation, maximizing equality of distribution (MED) is considered to be the objective of the WSPs. Subsequently, we have proposed simulated annealing (SA) based spectrum allocation solution that would maximize the objective. Our simulation result shows that the SA based solution outperforms existing solutions such as Shapley value and max-min fairness by wide margin as far as objective MED in concerned.