UPS Optimizes Its Air Network

United Parcel Service (UPS) started in 1907 as an American messenger service firm. Over the years, the company grew to where their services shifted from just private couriers and delivery to including retail package delivery in their services. By 1929, the company had adopted offering limited air delivery services for clients who needed their services between some west coast cities of the U.S. The company had developed to the extent of offering two-day delivery services in the main cities from coast to coast by mid-1950s. The company received a boost when they were allowed to provide their delivery services to all the 50 states by the federal authority. The company ever since has grown to the extent where they have expanded even to the rest of the world. The company became more complicated with more responsibilities as the schedules were tight. The services were done in this manner for the purpose of being more competitive in the delivery industry against their rival companies like FedEx Corporation. The organization, therefore, resolved in finding solutions to optimize their services in matters of spending their delivery by optimizing the air network Through the use of an application developed within the organization known as the Volume, Location, and Aircraft Network Optimizer (VOLCANO). This report will consider the problem the group was trying to solve, the results of the solution, the implementation process and the shortcomings of the solutions they adopted.

The problem and its context

The UPS airline delivery service operates in three different categories which include same day, next-day, and second-day deliveries. The company operates its network of all-points hub to ensure that the next-day deliveries are run smoothly. Additionally, the company has set strict deadlines for its customers to ensure they allow time for sorting the deliveries according to the destination before departure. Any late goods are to be delivered in the following scheduled pick up which is in the next day, and same day. The problem in this situation was on the next-day deliveries. The program runs under strict deadlines where the clients have to drop off the goods they need to be delivered between 6.00 pm and 8.00 pm. The company must deliver the goods to the destination by 10.30 am on the following business day which requires the organization to run its airports, aircraft, and hubs efficiently. The company has overflowing clientele where an average of more than 1 million packages is delivered in a single night from various branches of the organization. At every stop, packages are sorted regarding those being packaged to leave the hub for an outbound aircraft to those being unloaded to be loaded into trucks for deliveries at the destinations.

Planning for the plane routes, therefore, required lots of time and efforts as it involved numerous aspects. Some of the aspects include; sufficient speed to meet the deadline, restrictions on landing locations, maximum flying range and the cargo capacity of the aircraft. Manual planning presented a problem which was routes reversal where an aircraft could only travel one way which was the path to one airport or hub and back which would not utilize the aircraft to the maximum. The planning of this sort took up a large number of workforces from the planning and analyst sections of the organization. For example, three separate groups of planners worked in different next-day issues where they were tasked with identifying the most efficient route for the aircraft and package movements to ensure cost efficiency and timely deliveries. It would take the organization up to nine months to complete a single plan. The nine months period did not include analyzing the project and learning its workability. And with the broad range of data due to the dynamics of the organization, the planning process was becoming harder by the day. Planners and the analysts in the firm suffered the most as the task was on their hands. Finding solutions to all these problems was important for the company as it needed to maintain its timely deliveries, save on the cost by utilizing the aircraft’s capacity to the fullest and finding better routes to be able to save on the cost.

Analysis and approach

The organization focused the project on the next-day air system as planning with this category would allow the people involved to focus on a broad range of scenarios as the next-day network is more complex than the other two. By using the next day, it would also help create viable plans and to be able to generate an optimal plan for the next-day network. The project was based on optimization approach. The people involved used three methods to test and solve the problem. It was aimed at developing optimization methods. Through the optimization, the company would be able to save on the cost through determining the package flow and movements of the aircraft hence minimizing ownership.

One of the greatest challenges with the project was actualizing the theoretical concepts by presenting them to planners who were the ones tasked with approving them which is something that had never been done before. The team working on the project also faced problems in the process of actualizing it where the system itself presented a challenge in the network design components. The system itself ha to determine and match various aspects like minimum cost set for the route, package flows, speed, loading capacity and more. In other words, it was actualizing the system to optimize the full operation as it was supposed to function. The air design problem includes two decision variables which were integer-valued decision variables and would represent aircraft-routing decisions and continuous decisions variables which would represent package flow. Constrains presented the issue of tractability. The issue and much more showed the ineffectiveness of conventional means of trying to solve the network design problems. The team involved in the project abandoned resolving the problem with conventional ways and instead they adopted an interface design formulation whose linear programming relaxation was stronger. The group combined packages and aircraft routes to be a single integer-valued decision variable rather than using them separate as the previous proposition dictated. All this was in pursuit of optimal values of the cost of operation of the entire organization.

Other than technical challenges, the team faced organizational obstacles like resistance from the senior managers in the organization and the planning department. The management required the research team to justify proceeding with the project the planning group, on the other hand, saw the technological program as a threat as they had relied on their knowledge and skills for a very long time in the planning of the organization. The team also faced the challenge of receiving research funds. Although they were granted, the finances took very long to be processed. The team was, however, able to gain trust from the management through the monitoring, going through the documentations and seeing the progress made by the team. Through the participation of one planner, the team was able to gain favor from the rest of the executives in the firm as the single one on their side advocated for the changes they were a helped convince the rest on the workability of the project.

Results

The first results showed that conventional network design formulation needed more effort as compared to the combined variable formulation to get to the optimal value solution. The result was a seven percentage reduction on the operation cost. According to the organization’s saving, this would project to lots of millions for the company in just one year. Most of the savings come in the total number of aircraft in the operation as they would reduce and also on the ownership cost. As compared to the nine months period required by the planners to plan and additional time to analyze for the next-day air network, this technology would do all this planning and analysis in just two hours of running. The planners, however, are unable to adopt the VOLCANO fully but only use it to identify areas that need to be changed in their existing network plans. The use of the system is limited which means that the organization uses it as a tool to help in making organizational decisions, but they will soon realize the economic potential of the system.

Implementation

The program was implemented in 2000 and continued to be refined where it continues to cover more than just the next-day air network but includes the international delivery, three-day select, and second-day air. Since its implementation, VOLCANO has acted as more than just an optimization model. The system handles a variety of information which include; time windows of customer services, operating characteristics for aircraft, those of hubs and airports and volumes that move through the air network. The model has achieved in creating a precise timing that is required in scheduling the operations. The VOLCANO was integrated with the existing UPS databases and manipulated the data to be usable. An additional pre and post processing modules were built which was helpful in generating useful output for planners.  A commercial off-the-shelf integer and linear programming solver were used for the core analytical engine. The programming solver used was the ILOG CPLEX Callable Library. The C++ language was used to generate columns and structures for the integer program. The whole program runs on a Hewlett Packard N4000 machine where the HPUX 11 is the operating system.

VOLCANO has impacted on the organization in different ways. One of the impacts is the quantifiable cost saving. The impact is both short term and long term for the firm. The planners on the organization have used the VOLCANO since 2000 to make changes to the existing plans. The cost that has been reduced is the operating, ownership and even leasing cost. The operating expenses are crew and fuel; ownership is the cost of buying and maintaining the aircraft, and the leasing cost is one incurred in high season when the organization has to obtain additional aircraft through rental to increase their capacity. An example of a change implemented by the help of the VOLCANO is the shifting of two planes from the network to the poll spare aircraft in 2002 which reduced the cost of operation. The system has enabled to firm to lessen the number of aircraft by matching demand to the available capacity. The changes have helped reduce the cost with as much as $87 million, and it is estimated to save as much as $189 million in the coming decade.

Shortcoming

The VOLCANO just like any other technological advancement will render some individuals jobless. For example, the workforce that used to operate an aircraft that has stopped working will be forced to look for another job. The situation is also similar for the planners where a huge number of them will be obliged to go home with no job since the system has taken lots of work from their hands. The organization has become dependent on the system to help solve problems and is used as a consultant, neglecting knowledgeable people, therefore, rendering them jobless yet with the required skills.