In this study, we aim to design a controller (XCon) which optimizes energy-efficiency in the over-provisioned indoor 4G cellular networks for under-utilized eNodeBs. In this manner, we propose a novel energy efficiency coefficient (EEC) which covers conventional bits/joule energy efficiency metric and additionally contains parameters such as channel utilization and SINR. Moreover, we define an optimization problem which aims to maximize total EEC w.r.t network capacity and point-to-point delay (p2p). However, the introduction of the multi-parameter EEC increased the complexity of the optimization problem. In order to solve this problem, we propose a novel heuristic called "Two Phase Enhanced Branch and Bound" (TPEBB). This heuristic manipulates the network topology by closing underutilized eNodeBs during the first phase and forcing to offload an UE without closing any eNodeB during the second phase. Finally, we have compared the XCon with greedy and sleep-mode approaches under real-time and non real time traffic models in the evaluation section and shown that it provides moderate solution in terms of p2p delay and energy consumption.