Initialisation

Initialisation

Load data

Load m, the number of depots. Load K, the number of vehicles types. Load n, the number of customers.

Load distance matrix distanceMatrix into Matrix datatype with dimensions (n + m) x (n + m). Create depot matrix depotMatrix with dimensions m x (K + 1). Load customer matrix customerMatrix with dimension n x 2. Load vehicle matrix vehicleMatrix with dimension K x 4

Create empty solution

For each depot (m depots):

• Select row from depotMatrix
• Set inventory by calling Solution::addDepot(I_j)
• For each vehicle type:
  • Add n_jk vehicles by calling Depot::addVehicle(k) function

For each customer in customerMatrix:

• Create customer by calling Solution::addCustomer(demand, servicetime)

Construct initial solution

1. Sort customerMatrix on ascending demand.

2. Create depotCustomerAllocation (is a vector<vector<int>>). Create depotCustomerDemand.

3. Loop over customerMatrix:

   • Create leftoverDemand = customer.getDemand().
   • While leftoverDemand > 0
     • Find smallest distance to depots, say smallDepot.
     • Add customer.
   • end

4. Loop over depot (m depots):

   • Sort vehicles based on costs.
   • Create customerList and customerDropOff.
   • Loop over sorted vehicle list:
     • Create leftoverCapacity = vehicle.getCapacity()
     • while (leftoverCapacity > 0) do
       • Take new customer from depotCustomerAllocation (first time random).
       • Add customer to customerList and customerDropOff.
       • Update leftoverCapacity.
       • if (leftoverCapacity >= demand) then
         • Delete from depotCustomerAllocation.
       • else update depotCustomerDemand.
     • end
     • route = cheapestInsertion(customerList)
     • vehicle.setRoute(route, deliverQuantities)
     • Customers.addRoute(route, deliverQuantities)

cheapestInsertion(customerList)

1. Create route = {0, 0};
2. While customerList is not empty
   • For each customer in customerList
     • For each insertion option (N - 1):
       • Check if minimum
   • Update route by inserting cheapest.
   • Remove customer from customerList.
3. Ready to rambo.