I would like to factor in the CO2 emission in the Vehicle Routing Problem. The Arc Cost is not the distance anymore but the CO2 emission over each Arc
As a first order approximation, the CO2 emission of each arc can be modeled as a linear function of the current load and constant topographic characteristics of the arc (the slope, distance, altitude, ...) arc_cost(arc,load) = K(arc) + alpha(arc)*load
K and alpha are functions of the arc only (stateless) The load depends on the previous deliveries.
I haven't found how to model this with the ORTools Python API. Here is my current attempt : https://github.com/remisoulignac/scm_optim_problems/blob/main/SCM290-GreenVehicleRoutingProblem.ipynb
For now, I fall back on a two-level optimization approach: A classical VRP optimization model. I parameterize this model by several control points at which I impose a certain weight of goods to be delivered to the chosen route. A hyperparametric optimization level that will play with the different control points and optimize the overall real fuel consumption of the tour. Thank you for your help,
