<p>Heterogeneous Network (HetNet) is a cochannel deployment of macro base stations (mBSs) with very large antenna arrays as massive MIMO (Multiple Input Multiple Output) and a secondary tier of small cells (SCs) with a few antennas. A synchronized time division duplexing (TDD) protocol is utilized for signal transmission and reception in the network. The channel reciprocity is introduced in the proposed system that estimate the large-dimensional channels at the BSs and enable the implicit coordination between both tiers without the need to exchange user data or channel state information (CSI) over the backhaul. In this paper, a resource allocation with intra tier interference management system is proposed in the MIMO system. The resource allocation problem is formulated by nonlinear programming and the Q learning algorithm is introduced to overcome the optimization problem. In the spatial channel correlation with pilot contamination for massive MIMO network, the BSs and SCs suffer from dominant interference subspace in the network. In order to overcome the problem in the HetNet, transmit precoding and different receive combining schemes are proposed to reduce the inta-tier interference in the network. The Least Square (LS) channel estimator in uplink (UL) and downlink (DL) for pilot based channel estimation is implemented in the proposed system. The Simulation results for Spectral Efficiency (SE) in various receiver combination schemes with transmit precoding are obtained to reduce the intra-tier interference in a HetNet.</p>