An Adaptive Data-Driven Approach to Solve Real-World Vehicle Routing Problems in Logistics

Pseudocode 4

Route elimination.

	while(num_of_iterations --){
	vector<int> randPermutation = RandomPermutation (currentSolution.size());
	for(int i = 0; i < randPermutation.size(); i++){
	vector<Route^∗> routeCopy = currentSolution;
	for(int j = 0; j < routeCopy.size(); j++){
	routeCopy[j] = new Route(currentSolution[j]->customers, currentSolution[j]->routeVehicle, true);
	}
	double totalConsumption = 0;
	bool possible = true;
	for(int k = 0; k < routeCopy[randPermutation[i]]->customers.size(); k++){
	double minConsumption = CONSTANT;
	int route_number = −1, place = −1;
	for(int u = 0; u < routeCopy.size(); u++){
	if(u ! = randPermutation[i]){
	for(int = −1; =  = −1 || < routeCopy[u]->customers.size();++){
	if(routeCopy[u]->IsItpossibleToAddCustomer(routeCopy[randPermutation[i]]->customers[k],)
	&& minConsumption > routeCopy[u]->AdditionalConsumptionWithCustomerAtPlace
	(routeCopy[randPermutation[i]]->customers[k],)){
	minConsumption = routeCopy[u]->AdditionalConsumptionWithCustomerAtPlace
	(routeCopy[randPermutation[i]]->customers[k],);
	route_number = u;
	place = ;
	}
	}
	}
	}
	if(route_number =  = −1){
	possible = false;
	break;
	}
	routeCopy[route_number]->
	InsertRouteToThePlace(new Route(routeCopy[randPermutation[i]]->customers[k]), place);
	totalConsumption += minConsumption;
	}
	double savings = currentSolution[randPermutation[i]]->RouteCost();
	if(possible && (savings-totalConsumption > 0)){
	routeCopy.erase(routeCopy.begin() + randPermutation[i]);
	currentSolution = routeCopy;
	break;
	}
	}
	}