You need to  do Pareto optimization (cost/benfit) for specific speed at first, then variation of speed shall be modelled with the caused externalities. The value of externality can be added to the and change the values accordingly.

Motasem Saidan, thank you. Along your thoughts,i am thinking of optimising on the basis of weighted values that could be derived from say, the health impact that each of PMx, CO, HC, CO2, and NOx and so far i am establishing that PMx is the most dangerous emission and hence its weight is higher than the rest. The optimal speed mudt be a range that balances the cost