The number of mobile connected devices have increased exponentially worldwide. The services provided across mobile networks have also become increasingly demanding for higher network capacity. This has motivated the use of the heterogeneous network (HetNet) architecture. However, despite the low power consumption of small-scale base stations (BSs), their collective power consumption in dense HetNets is significant. The use of green energy to mitigate the power consumption of mobile networks is a trend on the rise. However, traditional association schemes under utilize green energy. Furthermore, in dense HetNets, there is an increased chance of users being on cell-edges and having degraded perceived service. Coordinated Multipoint (CoMP) association can aid in improving the service perceived by cell edge users. In this work, we propose a load balancing scheme that optimizes user latency and green energy utilization. The scheme allows for a fractional solution to the user association problem, enabling CoMP transmissions for cell-edge users. The proposed algorithm is a Quasi-Newton-based approach, which applies the Broyden-Fletcher-Goldfarb-Shanno (BFGS) method of approximating the inverse of Hessian matrices. Performance evaluation shows a reduction in latency of 79% and a reduction of power consumption by 99% in comparison to conventional schemes.