Using OR-tools documentation I have combined CVRP with PD and TW. In this problem, since the pickup at nodes exceeds the vehicle capacity, nodes were duplicated to keep the pickups equal or less than the vehicle capacity and accordingly the O-D matrix is modified. The list of P-D pairs has been taken. The pickup node has a positive demand and the delivery node has a negative demand. Each node is associated with the service time window.
There are several things which I am unable to figure out and would require help from the community and experts in this domain. I am listing the doubts I have as of now
1. Number of Vehicles
The Approximate number of vehicles is given as input, and different search strategies give different optimal solutions on every run. Solutions either have more vehicles with less objective value or less vehicles with more objective value. I want the minimum number of vehicles required to complete the VRPDTW, so can we use fixed cost for using a vehicle and how it can be coded?
2. Local Search Strategy and Metaheuristics
The solutions obtained differs for different combinations of strategies. I want to understand the parameters and the method to evaluate different solution strategies.
3. Time Limit
How can we understand what time limit needs to be provided to arrive at a solution which cannot be improved for different size of the problem?
4. Objective Value
Could you please explain how the objective value is calculated?
Please explain the use of this function and where should i put it in the code