In this paper, we study Base Station (BS) switching-off and offloading for next generation 5G heterogeneous (macro/femto) networks supplied with hybrid energy sources. This type of network will form the basis of the high-data rate energy- efficient cellular networks in the years to come. A novel generalized multimetric algorithm is presented. Our proposal is conceived to operate in highly heterogeneous Radio Access Network (RAN) environments, as expected for 5G, where BSs with different characteristics of coverage, radio resources and power consumption coexist. The approach uses a set of metrics with a modifiable priority hierarchy in order to filter, sort and select the BS neighbors, which receive traffic during redistribution and offloading of the BSs to be put into sleep mode. In our analysis, we study the impact of BS power model trends for active, idle and sleep modes on the BS switching-off. We highlight how the continuous evolution of BS components and the introduction of renewable energy technologies play a significant role to be considered in the decision making. The multimetric approach proposed makes it possible to define and accomplish defined network performance goals by adding specific emphasis on aspects like QoS, energy savings or green equipment utilization.