This study aims to develop two optimization models that provide a realistic representation of interactions between selfish users and C-RAN networks by integrating demand response mechanisms, heterogeneous incentives, and quasi-linear utility functions. The first model is designed to formulate selfish user behavior in a C-RAN system based on utility structures and incentive schemes, while the second model extends this framework through the incorporation of a bundling scheme to evaluate improvements in network efficiency and user utility. The research methodology involves collecting and formulating 30 days of traffic data, defining relevant model parameters and variables, constructing the optimization models, and solving them using LINGO 13.0 under three pricing schemes: usage-based, flat-fee, and two-part tariff. The results reveal that the flat-fee pricing scheme with bundling in Case 2 provides the most efficient configuration, achieving the highest objective value of and the lowest iteration count. These findings demonstrate that integrating bundling strategies into C-RAN pricing models can enhance network efficiency, improve bandwidth utility, and increase ISP revenue.