7 Challenges        

Most supply chains have a lot of data but still can’t decipher good information. Adding to the data complexities are the timely availability of the same. One example that we’ll consider here is Vehicle Routing and Scheduling. This area within supply chain is the most vulnerable resulting due to data unavailability or delayed information.

The vehicle routing problem faced in practice involves many restrictions on the routes that delivery vehicles can follow and we consider some of the more common restrictions below. We can classify these restrictions to a certain extent as relating either to the vehicles or to the customers (shippers).

  Each vehicle has a limit (capacity - usually weight and/or volume) on the goods carried.  
  Each vehicle has a total working time from departure to arrival back at the depot, typically to comply with legal restrictions on driver working hours.  
  Each vehicle has a time period within which it must leave the depot, typically to ensure that space is available for incoming vehicles to re-supply the depot.  
  Each vehicle has a number of time periods during which it does nothing (driver rest periods).  
  Each vehicle has a cost associated with its use for deliveries.
  Each customer has a certain quantity which has to be delivered. Sometimes this quantity is known exactly (the deterministic case) and sometimes known with a degree of uncertainty (the stochastic case).    
  Each customer has a number of time periods during which delivery (and/or collection) can occur (time windows).    
  Each customer has an associated visit time (drop time).    
  Each customer has a set of vehicles which cannot be used for delivery (access restrictions)    
  Each customer has a priority for delivery (if the vehicles cannot deliver to all the customers). Typically this might happen due to driver/vehicle unavailability or due to poor weather conditions dramatically reducing vehicle speeds    
  Each customer may accept split visits (a delivery/collection by more than one vehicle) or not.
Other Factors:
  Multiple trips by the same vehicle on a single day, where the vehicle returns to the depot and then goes out again.  
  Trips by the same vehicle longer than one day.  
  Compartmentalized vehicles with many different types of product to deliver.  
  More than one depot, where vehicles can start/visit/end at different  

In the design of vehicle routes to meet the above requirements there are a number of objectives that could be adopted. Three basic objectives can be distinguished:

  Minimize the number of vehicles used  
  Minimize the total distance (or time) traveled  
  Minimize some combination of number of vehicles used and total distance (or time) traveled

The cost of vehicles in the vehicle fleet is often regarded as a fixed cost so that the first objective above corresponds to the minimization of fixed costs; the second objective above corresponds to the minimization of the running (or variable) costs; and the third objective above corresponds to the minimization of total cost (the sum of fixed plus variable cost).

A Mathematical Model for Vehicle Routing and Scheduling with Time Windows:

PCI delivers optimization models for supply chains to mange cost-to-serve and meet shipper’s on-time delivery performance.