Cognitive Vehicular Network for Optimal Resource Allocation Using MINLP

Abstract: Dedicated Short Range Communication (DSRC) band is incapable to remove increasing demand on wireless traffic in vehicular network. The TV white Space band by FCC for cognitive access provides additional bandwidth to solve the DCRS spectrum problem. However, create a challenging environment for portable (e.g., vehicular) and fixed (e.g., IEEE 802.22) network which is FCC required portable device to use significantly lower transmitting power than fixed device. In this paper, first formulate the Mixed-Integer Non Linear Programming (MINLP) program, to which three algorithms are refined. The first algorithm converts the MINLP to a convex problem and gives the near-optimal solution to the initial MINLP. The other two algorithms, first convert the MINLP into an Integer Programming (IP) problem. Then, solve the linear program relaxation of the IP and obtain fractional solution. Consequently, two rounding algorithms are developed to round the fractional solution based on the column-sparse packing and dependent rounding techniques. In conclusion, compare the performance of the proposed algorithms with the optimal MINLP solver.

Keywords: Channel Allocation, Cognitive Vehicular Network, Linear Programming, Submodular Set Function