Performance Analysis of Optimization Algorithms for a Cross-Layer Shadowing Recover Scheme

Abstract

Optical attocell networks based on visible light communication (VLC) are cost-efficient solutions which complement the spectrum shortage of radio-frequency(RF)-based indoor wireless networks. However, communication using visible light is vulnerable to shadowing effect caused by random pedestrians as visible light cannot penetrate opaque objects. A cross-layer resilience scheme RASP (routing, load assignment, subcarrier selection, power allocation) is proposed to address the shadowing effect. Network performance using different combinations of cross-layer control policies, i.e., algorithms and allocation methods, is evaluated and compared by RASP. An optimization algorithm is used to find the optimal/near-optimal combination of RASP control policies. In order to find the proper optimization algorithm for RASP, we investigate and compare the performance of different algorithms, including partial enumeration and neighbor-searching algorithms like hill-climbing and Tabu-search. An improved Tabu-search algorithm is proposed to reduce the time-complexity.