Font size

Phase 1. Import Scenario and Run the Experiment

We shall start with obtaining the data that will serve as the basis for our experiment. You may either provide your own data or you can instantly import the scenario. Either way, you should have a supply chain scenario containing the following:

  • Customers
  • Demand
  • Distribution center(s) located in the optimal places
  • Inventory
  • Properly configured sourcing and paths
  • Vehicle type(s)

The numbered lists in tutorials are actually checklists. Click the numbers to save your progress!

Download and import the scenario

  1. Download the scenario. No internet connection is required, since the scenario is supplied with anyLogistix.
  2. Import the downloaded scenario. The GIS map will appear showing the content of the imported scenario.
  3. Explore the connections by clicking the Show filters button to open the filter options, and then clicking the Show connections button.

The imported scenario contains a supply chain with 32 Lidl stores and one distribution center on the territory of Bulgaria.

Let us observe the data the experiment will work with. Primarily we are interested in the Paths and the Sourcing tables, and Vehicle Types.

Observe the data of the scenario

  1. Navigate to the Paths table. It contains 1 record, allowing all connections between the objects of our supply chain. If you need to edit this record, make sure that the resulting record(s) (with Fixed delivery cost or Distance based cost calculation formulas only) will allow a vehicle to go from:
    • Site to Customer
    • Customer to Customer
    • Customer to Site
  2. Now switch to the Sourcing table, which defines the sourcing of products within this supply chain. The Sources column contains the Gabrovo DC, which will be servicing customers that are defined by the Lidl Stores group in the Delivery Destination column.
  3. Lastly, open the Vehicle Types table. You will see the vehicle that we will be using. It is a standard 40 m3 truck, that is specified in the Paths table, and in the experiment parameters.

Configure parameters of the CTO experiment

  1. Navigate to the experiments section and click Capacitated TO experiment.

    You will be taken to the experiment's view with its settings, where:

    • Number of shipments — the number of shipments you would like to send within the Experiment duration period. We will have 8 shipments within 2 months, i.e., 1 shipment a week.
    • Vehicle types — the type of vehicle delivering the products. The selected vehicle type(s) must correspond to the vehicles defined in the Paths table.
    • Travel segment limit — the maximum remoteness of the network objects from each other in the specified Distance unit.
    • Returning segment limit — the maximum remoteness of the last customer of the route from the DC that the vehicle set off.
    • Min vehicle load ratio for direct shipments — specifies the minimum amount of products a vehicle can carry as a direct shipment (shipment made to one customer only). In our case, a customer must order 28 m3 of products or more to have a direct shipment. The value is calculated as vehicle capacity (40 m3) multiplied by the parameter's value (0.7).

  2. Click Run in the toolbar of the CTO experiment.
    The results will be available in the Result sub-item of the CTO experiment tree branch.

Let us analyze the received data.

Analyze the received results

  1. Navigate to the Optimization results page below the experiment view. The results contain all the possible routes with the lowest expenses. There are 8 sets of routes, one for each shipment. Each set includes only the routes that let a vehicle visit every customer by the specified vehicle.

  2. Expand the borders of the Destinations column to observe the order in which customers are visited within each route.
  3. Filter the results per shipment. Type 01-07 (the date of the first shipment) into the filter field below the Aggregation Period column name.

    Scenario duration period is 2 months. The Number of shipments parameter is set to 8 shipments, which results in weekly shipments, hence the date of the first shipment in the results.

  4. Now click a result record. The GIS map will appear. In our case it comprises 8 customers: Gabrovo, Ruse, Shumen, Dobrich, Varna, Burgas, Stara Zagora, Kazanlak.

    Capacitated TO experiment considers actual roads. The GIS map, however, depicts the results with straight lines by default. Disable the Show Straight toggle button in the Optimization results page to depict actual routes.

  5. In the same way you may observe the rest of the offered routes on the GIS map.
  6. Now switch to the Solution Type column. It offers 2 types of solution for routes of our supply chain:
    • Optimal — optimal solution in terms of costs and the provided data.
    • Direct — the shipment made directly to a customer whose weekly demand exceeds the capacity of our vehicle (40 m3). The rest of this customer's demand is used by the experiment to design the delivery network.
  7. Navigate to the Generated Paths page of the experiment results to observe transportation cost per route. The table contains dozens of records, showing routes per each shipment.

  8. Group the records by shipment by clicking the Shipment column title and dragging it to the Grouping area.
    The first line contains summarized data on all the routes from the first shipment, lines 2-5 show data per each route of this shipment.

  9. Check the Solution Type column. As you can see, we have 1 direct shipments and 3 optimal routes within the first period.

  10. Calculate the sum of all the routes. In our case the total cost of the offered routes constitutes $587.312.

  11. For more details you can navigate to the Generated Path Segments page, which contains details on each segment of every route.

That's it we have completed the first phase of this tutorial. We imported the scenario, created cost-efficient routes, and analyzed the results. In the next phase we will try to decrease the cost of running the supply chain by adjusting the experiment parameters.

How can we improve this article?