Import distance matrix from CSV

[Bryan Salazar López]

Hi guys, I'm new to this group and I really don't know how I've progressed in building my model, since I'm an Industrial Engineer and I'm not proficient in Python.

I have managed to build the script that imports the distance matrix from "Distance Matrix API" and then operates them by multiplying matrices and scalars, transforming a matrix of distances and a matrix of times, into a matrix resulting in costs.

This cost matrix I can export in multiple formats: txt / csv, and of course it is a variable from my script.

I have not succeeded in getting the script that contains the Or-tools model to import the matrix from the script that generates the cost matrix. However, as I have mentioned, I have the csv and txt files.

Can you help me in the structure of the model in Python that imports a csv file as the distance matrix of the Or-tools model?

I would greatly appreciate your collaboration, this optimization module is part of a project to analyze the process of generating biodiesel from used cooking oil.

[Madushan Fernando]

Hi Bryan, 

I used the following python code to import data from CSV and create the nested matrix. Then input that matrix as distance matrix.

import pandas

##Import CSV##

pandas.set_option('expand_frame_repr', False)

df = pandas.read_csv(r'C:\Users\User\Desktop\My working\Distance.csv')

print(df)

##Dataframe to list##

Matrix= df["Distance"].tolist() 

print(Matrix)

##Convert into nested List##

i=0

new_list=[]

while i<len(Matrix):

  new_list.append(Matrix[i:i+12])

  i+=12

print(new_list)

--
You received this message because you are subscribed to the Google Groups "or-tools-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to or-tools-discu...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/or-tools-discuss/78eaae49-fac1-49cc-b3c9-2446eaa568den%40googlegroups.com.

[Bryan Salazar López]

I attach the scripts that I use to call and operate the distances (D_matrix) and the optimization script (prueba_penalties)

[Bryan Salazar López]

Hello, I appreciate your answer, I do not know if you could observe my .csv

I don't know if your code applies to my format, since the matrix is practically built

In fact, I'm going to show you the output I have from my D_matrix.py file:

This output is named as variable F in my script.

How can I invoke it from the optimizer script? I see that it is in the precise format that Or-tools requires

Or, How can I import my .csv in the format I attached above?

[Madushan Fernando]

I imported data from the following CSV and it contains a distances data frame.

To view this discussion on the web visit https://groups.google.com/d/msgid/or-tools-discuss/b57740cd-6ac3-40c4-9189-2f72cfa90bafn%40googlegroups.com.

[Bryan Salazar López]

Now I understand your code a bit. No, I already have my csv in matrix form

An n x m matrix, where n = m

[sirinebe...@gmail.com]

Which is the feasible Dataframe or dictionary? I import data from csv file after  i convert to dictionary data. it  read correctly

 But,  the code of TSP   displays the distance = 0

[Bryan Salazar López]

It has worked like this for me (VRP):

def matrix_distance():

  matrix = np.genfromtxt('distance_matrix.csv', delimiter=';')

    

  return matrix

def create_data_model():

    """Stores the data for the problem."""

    data['distance_matrix'] = matrix_distance()

    data['vehicle_capacities'] = vehiculos_capacidad

    vehiculos_numero = len(data['vehicle_capacities'])

    data['num_vehicles'] = vehiculos_numero

    data['penalties'] = penalty

    data['depot'] = 0

    return data

    def distance_callback(from_index, to_index):

        # Convert from routing variable Index to distance matrix NodeIndex.

        from_node = manager.IndexToNode(from_index)

        to_node = manager.IndexToNode(to_index)

        return data['distance_matrix'][from_node][to_node]

    distance_callback_index = routing.RegisterTransitCallback(distance_callback)

[Sirine Belguith]

Thanks Bryan.

it returns the distance of the route  and objectives =0.

Below is my code:

def matrix_distance():

    matrix = np.genfromtxt('vehicle (2).csv', delimiter=';')

    

    return matrix

def create_data_model():

    """Stores the data for the problem."""

    data = {}

    data['distance_matrix'] = matrix_distance()

    data['num_vehicles'] = 1

    data['depot'] = 0

    return data

def print_solution(data, manager, routing, solution):

    """Prints solution on console."""

    print(f'Objective: {solution.ObjectiveValue()}')

    

    max_route_distance = 0

    print(len(data['distance_matrix']))

    for vehicle_id in range(data['num_vehicles']):

        index = routing.Start(vehicle_id)

        plan_output = 'Route for vehicle {}:\n'.format(vehicle_id)

        route_distance = 0

        while not routing.IsEnd(index):

            plan_output += ' {} -> '.format(manager.IndexToNode(index))

            previous_index = index

            index = solution.Value(routing.NextVar(index))

            route_distance += routing.GetArcCostForVehicle(

                previous_index, index, vehicle_id)

        plan_output += '{}\n'.format(manager.IndexToNode(index))

        plan_output += 'Distance of the route: {}m\n'.format(route_distance)

        print(plan_output)

        #{print(index)

        max_route_distance = max(route_distance, max_route_distance)

    print('Maximum of the route distances: {}m'.format(max_route_distance))

def main():

    """Entry point of the program."""

    # Instantiate the data problem.

    data = create_data_model()

    # Create the routing index manager.

    manager = pywrapcp.RoutingIndexManager(len(data['distance_matrix']),

                                           data['num_vehicles'], data['depot'])

    # Create Routing Model.

    routing = pywrapcp.RoutingModel(manager)

    # Create and register a transit callback.

    def distance_callback(from_index, to_index):

        """Returns the distance between the two nodes."""

        # Convert from routing variable Index to distance matrix NodeIndex.

        from_node = manager.IndexToNode(from_index)

        to_node = manager.IndexToNode(to_index)

        return data['distance_matrix'][from_node][to_node]

    transit_callback_index = routing.RegisterTransitCallback(distance_callback)

    # Define cost of each arc.

    routing.SetArcCostEvaluatorOfAllVehicles(transit_callback_index)

    # Add Distance constraint.

    dimension_name = 'Distance'

    routing.AddDimension(

        transit_callback_index,

        0,  # no slack

        3000,  # vehicle maximum travel distance

        True,  # start cumul to zero

        dimension_name)

    distance_dimension = routing.GetDimensionOrDie(dimension_name)

    distance_dimension.SetGlobalSpanCostCoefficient(100)

    # Setting first solution heuristic.

    search_parameters = pywrapcp.DefaultRoutingSearchParameters()

    search_parameters.first_solution_strategy = (

        routing_enums_pb2.FirstSolutionStrategy.PATH_CHEAPEST_ARC)

    # Solve the problem.

    solution = routing.SolveWithParameters(search_parameters)

    # Print solution on console.

    if solution:

        print_solution(data, manager, routing, solution)

    else:

        print('No solution found !')

if __name__ == '__main__':

    main()

Garanti sans virus. www.avast.com

To view this discussion on the web visit https://groups.google.com/d/msgid/or-tools-discuss/51ee3708-71c4-42f5-889b-8d59953a2503n%40googlegroups.com.