In this thesis, we study the stream selection based interference alignment (IA) algorithms, which can provide large multiplexing gain, to deal with the interference in the heterogeneous networks. Firstly, different deployment scenarios for the pico cells are investigated assuming perfect channel state information (CSI) at the transmitters. Two different stream selection IA algorithms are proposed for fully and partially connected interference networks and selecting at least one stream is guaranteed for each user. A stream sequence is selected among a predetermined set of sequences that mostly contribute to the sum-rate while performing an exhaustive search. In the proposed algorithms, the complexity of the exhaustive search is significantly decreased while keeping the performance relatively close. After selecting a stream, the interference generated between the selected and the unselected streams is aligned by orthogonal projections. Then, the influence of the imperfect CSI on the proposed algorithms is analyzed and it is observed that the intra-stream interference causes a significant degradation in the performance due to the quantization error. Therefore, we propose an algorithm for the limited feedback scheme. Finally, adaptive bit allocation schemes are presented to maximize the overall capacity for all the proposed algorithms. The performance evaluations are carried out considering different scenarios with different number and placements of pico cells. It is shown that the proposed algorithm for the limited feedback is more robust to channel imperfections compared to the existing IA algorithms. The presented bit allocation schemes improve the performances of the algorithms compared to the equal bit allocation.