Modeling periodic scheduling and balancing

Periodic scheduling and balancing selects from a number of potential delivery schedules, selecting the optimal delivery schedule based on several factors. As well as pool site balancing, routes are formed, considering the potential location schedules such that transportation cost is minimized. If you have included inventory holding cost and capacity at the customer locations, the frequency of visits is optimized and the holding cost is balanced against the transportation cost, while ensuring that stock outs do not occur.

The periodic scheduling and balancing problem type uses standard historic shipment data and specifies schedules per location. The shipment data is populated in the Transportation Daily Orders table, where you define the daily consumption at a destination along with the source. Define the orders within an offset value in the periodic schedule. For example, if the Start Day Offset Value is 1 and the End Day Offset Value is 3, the order amount must be delivered on or before each of the first three days of the fixed, often weekly or bi-weekly schedule. Transportation Daily Orders are not product-specific, they are the cumulative orders for the source-destination pairs.

You can alternately define shipment data in the Shipments table, however you then need to define actual shipments windows based on calendar days as well as provide the visit frequency for each location.

You populate the Periodic Schedule table to set the inventory holding cost and capacity. You can use this table to specify the frequency of deliveries and identify days when orders must or must not be delivered.

Periodic optimization uses the same constraints for routes as standard optimization. In addition, you can specify the Minimum Load Per Day and/or Maximum Load Per Day constraints in Transportation Optimization options. These constraints help perform load balancing across the model horizon. Unlike the standard optimization, the Available Units for asset availability is per day, not for the entire model horizon.

The periodic scheduling and balancing problem looks to find a visit sequence for each location and create routes for each day with a goal of balancing holding cost against a minimized transportation cost. There is no restriction that deliveries made 2 times per week are to be combined on the same route as other twice weekly deliveries. Some of the twice weekly deliveries could be paired with deliveries made three times per week if that is optimal from a pool site balancing and routing perspective.

You can select "Periodic Optimization" as the Problem Type on the Advanced tab in Transportation Optimization options, or set the Problem Type to "Auto-Detect". Transportation Optimization will run periodic optimization based on the inputs you have defined.

When periodic scheduling and balancing is run, it uses two general phases. It first builds an initial solution, then performs an improvement phase.

During the initial solution phase, the solver:

• Finds a visit sequence for each location, by routing the shipments then comparing this solution to the visit sequences. Once it finds a routing solution that does not violate the visit sequence, it assigns this visit sequence to the initial solution.
• Generates the routes using the visit sequence selected in the first step.

In the improvement phase, the solver looks to improve the solution by considering alternate visit sequences and then by moving or exchanging shipments between routes on the same day.

You can perform Fleet Optimization by selecting the Optimize Fleet Size option on the General tab of Transportation Optimization options.

You can review the results of periodic optimization using the following output tables:

• Periodic Optimization Summary - For each source-destination combination, this table provides the frequency of visits, the days on which visits occur, the inventory holding cost, allocated transportation cost and amount of delivered product.
• Periodic Optimization Inventory Details - For each source-destination-date combination, this table provides inventory holding cost, allocated transportation cost, and the amount of product received and consumed. This table applies to Transportation Daily Orders.
• Delivered Shipments - When using Shipments or Transportation Daily Orders as input, this table provides details about the shipments that were included in the periodic optimization, include their allocated route cost and pickup and delivery times.