Real-estate owners pay for district heating in the following two ways: consumption, measured in kWh; and demand, measured in kW. This means that the cost for utilizing district heating for a building is a function of both the total energy (consumption) utilized by the system and of the maximum power that the heating system has utilized (demand). This is a common economical model that the district heating suppliers employ in order to force the real estate owners to keep their maximum power on lower levels, and the term that is often used is demand charge tariffs. The reason for including demand charge tariffs is that the district heating supplier want to reduce the peak demands that occur in the system on specific occasions. The traditional approach for a real estate owner to perform control of a building’s heating system is by simple open-loop control that puts out a specified effect or feed temperature which only depends on the current outer temperature. A smarter way of controlling the system is by utilizing model predictive control (MPC) algorithms that models the building’s behavior and controls the heating systems with specified objectives and constraints. With MPC one can employ the true economic model and hence try to minimize the cost (consumption and demand) of the real estate owner when controlling the heating system. In this work we analyze how MPC algorithms are affected by the economical models that the heating supply companies utilize and demonstrate that it is possible to reduce the heating system cost for the real estate owner by introducing MPC problems that have both the consumption and the demand cost as objectives. We analyze what happens when one assumes that the demand charge tariff (the cost for the maximum demand) can be shared between a population of buildings instead of being charged per individual building and demonstrate that such an economic model can be beneficial both for the real estate owner as well as the district heating supplier since the coupled problem can plan and reduce the overall peak load of the district heating system. All evaluations and simulations are performed using real world data 57 Smart Energy System analyses from a population of buildings in Sweden where we currently are controlling the heating systems with our developed algorithms.