- governmental restrictions on driving time and rest periods
- customer preferences among carriers
- contractual minimum volume agreements with carriers
- carrier capacity limitations
- loading and unloading space constraints
The Transportation Planning Problem
A typical manufacturing company arranges and purchases transportation from a variety of carriers - from small package transporters like FedEx and UPS to full truckload operators.Some types of carriers - typically the small package (parcel) and less-than-truckload (or LTL) carriers - charge by unit of weight, with the rates varying by the distance covered.Full truckload carriers, however, as the name implies, sell a full trailer's worth of transportation on the desired itinerary.They typically charge only by the distance traveled, regardless of whether the truck is completely full or just carrying a small amount of goods.Usually, a well-utilized full truckload represents the most cost-efficient means of domestic transportation.Thus, the cost-minimization challenge faced by transportation managers involves not just picking the lowest-cost carrier for each shipment but, also, finding ways to combine smaller shipments into full truckloads.
The most obvious way to do this involves merely aggregating shipments that travel between the same origin and destination. Logisticians, however, really earn their pay by applying more complicated consolidation strategies.For instance, a full truckload carrier can execute a multi-stop load.Such a trip, as the term indicates, visits multiple origins and/or destinations in order to get multiple shipments onto the same truck.Another consolidation strategy, known as pooling or zone skipping, involves the use of hub facilities to consolidate shipments.For example, a parcel or LTL carrier might pick up small shipments and bring them to a nearby hub, where they are combined into a full truckload that traverses the bulk of the journey.The pooling strategy may utilize local pickups and consolidation (as in the example just given); deconsolidation and local deliveries; or both.In any case, pooling and multi-stop trips, if done right, offer great savings opportunities.The huge number of possible ways to combine the hundreds or thousands of daily shipments into truckloads, though, combined with the confusing array of constraints, makes it very difficult to identify optimal (or even near-optimal) arrangements without automated assistance.
Uses of Geographic Information in Oracle Transportation Planning
A key part of a transportation management system's effectiveness stems from the accuracy and precision with which it models the logistics environment.Naturally, as transportation involves moving goods from one location to another, spatial modeling constitutes an important element thereof.Also, an effective TMS user interface demands the ability to present huge amounts of complex data in a manner that the user can easily comprehend.The geographical paradigm represents an outstanding way to achieve this.
Spatial and Distance Modeling
A great variety of locations appear in the typical transportation planning problem.Of course, the plants, distribution centers, etc., of the company planning its own transportation must appear. So, too, must its suppliers' and customers' facilities, as these constitute, respectively, the origins of inbound and the destinations of outbound shipments.Finally, carrier-owned consolidation and deconsolidation sites (pools) form a significant element of the transportation picture.As an initial step, geocoding is employed to store all of these locations.
An important use of the location data involves distances and transit times between points.A TMS needs accurate transit times, of course, to achieve on-time deliveries.In some cases, distances are used in the calculation of these transit times.In addition, the TMS always needs accurate distances to determine the costs of truckload transportation correctly, as this is usually priced per unit of distance.Obviously, the system cannot select cost-optimal solutions without visibility to true costs.
Oracle Transportation Planning allows users to determine distances and transit times based either on an actual highway network or on a simpler approximation.The network approach, of course, requires extensive data on the links in the network, with a distance and effective traversal time for each.For users wishing to avoid acquisition of this data, a less-demanding alternative can be employed.This involves, first, calculation of a Euclidean distance between points, based on the points' geocode data.Next, the system applies a simple scale factor. Empirically, in the absence of significant geographical barriers (e.g., away from the Great Lakes), this approach produces more accurate results than might typically be expected from such a simple method.Finally, the program derives an approximate transit time from the estimated distance and expected speed factors.
Map-Based User Interface
A typical manufacturing or distribution company may process thousands of orders each day, each of which must be transported to, from, or between the company's facilities.These orders may be grouped into hundreds of truckload or other shipments.This huge volume of data elements makes it difficult for a transportation planner to quickly absorb the configuration and quality of a plan to transport one day's orders.Presenting the plan graphically on a map can make it much easier for the planner to digest.Seeing the plan's layout geographically allows the user to determine quickly if everything "makes sense" or not.For example, an overly circuitous multi-stop truckload stands out quite obviously on a map.Even when everything appears to be in order, the map representation provides an intuitive "index" for the user to drill in to more detailed information.Oracle Transportation Planning, using the Oracle AS MapViewer, includes interactive access through a map to all of the supporting transportation plan data (Figure 1). For instance, the user can highlight the line representing a truckload on the map and then click through to screens giving all the details of the truck's route, schedule, load, costs, etc.
The interactive nature of the map takes it beyond serving as a mere graphical representation or access tool for the transportation plan.The map also serves as a powerful problem-solving tool.For example, the many constraints that can be modeled may result in a truckload that combines several small shipments, while representing the lowest-cost option, still having a utilization level below target.In such circumstances, truly determining the "best" option may require some human judgment, evaluating whether the cost impact of a constraint may actually justify its situational override.The user may want to select some additional candidate small shipments; identify which constraints would be violated by adding them to the underutilized truckload (and how severely); and evaluate the cost savings from doing so.Then, the user can apply their experience and expertise to the decision of whether or not to fill out the truckload further.But how can the candidate small shipments be generated? No simple rules or easy-to-compute measures exist to narrow this search down.Any human transportation planner, however, can easily pick out "geographically compatible" truckloads and shipments when viewing them on a map.To take advantage of human input in such circumstances, the user can incorporate problem-solving workflows that utilize the map directly into the process.For instance, to solve the problem just described, a logistician can put the possible candidate shipments on the map with the underutilized truckload; select candidates for further evaluation directly from the map; and perform the required analysis without leaving the map framework.In this manner, the map viewer is used to facilitate not only the communication of the plan but also its improvement.
Enterprise transportation optimization needs to leverage geospatial information at several levels.Existing applications for this purpose often produce automated "optimal" solutions that are not really useable because they fail to model the logistics scenario completely.As a result, their users must spend hours wrestling with their output, trying to solve problems caused by the failure to model important considerations.Rich modeling of the planning problem - including an extensive constraint set and solid geographical data - creates optimization output that is much more executable.Furthermore, interactive problem-solving tools, including map-based tools, speed any required post-optimization analysis.In this manner, location-based modeling is an important part of advancing the state of the art in logistics management.