Competitive Spectrum Allocation with Interference

With the advent of 5G network slicing, demands for spectrum are expected to originate from a variety of services with different requirements. Given the ever-increasing scarcity of spectrum, a key engineering question is how it should be allocated to meet the needs of these services. While previous work has studied spectrum auction design or interference management techniques, little work has been done to combine these two areas for network slicing. We address this by formulating a weighted Vickrey-Clarke-Groves (VCG) mechanism which uses a bias term to penalize “spectrum pollution”. After stating the resource allocation model, we discuss the design and characteristics of the bias term and analyze important economic properties of the mechanism. Our simulation results show a significant improvement in communication performance compared to the vanilla VCG mechanism, and also show that interference-awareness can sometimes improve revenues.