Multi-commodity Flow Problem for Multi-hop Wireless Networks with Realistic Smart Antenna Model

Utilizing smart antennas in multi-hop wireless networks can boost their performance. However, most of the work done in the context of wireless adhoc and sensor networking assume ideal or over-simplistic antenna patterns that overrate the performance benefits. In this paper, we compute the optimal throughput of multi-hop wireless networks with realistic smart antenna model. Our goal is to evaluate the network performance degradation using real smart antenna as compared to using the ideal model. We derive a generic interference model that can accommodate any antenna radiation pattern. We formulate the problem as a multi-commodity flow problem with novel interference constraints derived from our generic interference model. Our numerical results show that using a real-world smart antenna in a dense network results in up to 55% degradation in the throughput as compared to the case of ideal flat-topped antenna, whereas the throughput degradation is as much as 37% in a sparse network.