Optimize dependent variable (optimize preheating signal to reduce cost/energy consumption of heatpump)

I have applied a simple preheating application to the matlab/simulink House Heating System (Matlab Command "ssc_house_heating_system"), and I want to estimate the optimal amount of preheating to minimize the total energy consumption/cost of the heat pump. I integrated an ADD block that adds additional "on" signals to the on/off timeseries generated by the thermostat. The additional "on" signals were input manually to a separate timeseries table, 1 is on 0 is off; this is the table that i want automatically optimize.

The results showed reduced cost (reduced overal energy consumption by heatpump due to preheating) by manually appliying one instance of preheating before the thermostat sent it first "on" signal, as a proof of concept. However, I want to optimize the additional "on" signals in the dataframe added to the on/off signals from the thermostat to minimize the cost.

What is the best approach to achieve this automatically with matlab when the outdoor temperature is known? This model has a known and predifined outdoor temperature ("daily temp variation") which affects the thermal network of the house.

This is what the preheating table look with with manual input of "on" signals (1's)

The heating results show indoor temperature (purple) and outdoor temperature (yellow). The total accumulate cost in the second graph is what I want to minimize with preheating.