Bright prospects for mailmen: how algorithms render logistics more efficient

Nowadays, customers who order their products online or have gotten used to well-stocked shelves in the supermarkets tend to forget that these conveniences always take a great deal of planning and logistics. “The greatest possible efficiency of such processes is one of the major determining factors for the companies’ competitiveness”, explains Prof. Dr. Heinrich Kuhn, Chair of Supply Chain Management & Operations at the KU’s Ingolstadt School of Management. Most recently, Prof. Dr. Kuhn and his team hosted a nationwide four-day workshop at the castle Schloss Hirschberg in Beilngries for doctoral and habilitation candidates and junior professors focusing on current topics in the field of quantitative business administration in their scientific work. The early career researchers were particularly interested in analysis, structuring and optimization of processes in procurement, production, provision of services and distribution. “They develop mathematic models and algorithms to increase companies’ competitiveness and, in particular, enhance customer benefit. Furthermore, sustainability, social ethics and ecological aspects become increasingly important in this field”, says Kuhn.

On the occasion of the conference, more than 100 participants from over 30 renowned academic chairs all over Germany presented their current research projects – among them also Prof. Kuhn’s doctoral candidates who cooperate amongst others with grocers, retailers, industry companies and logistics service providers in the context of their projects. Tobias Potoczki, for example, investigates structures of distribution networks for trading companies that transfer products from the producer to the store. “The easiest process is a direct delivery concept in which the producer delivers the goods to the stores. As soon as the number of delivery companies, different products and stores of a company increases, the process becomes more and more inefficient, as transport time and costs can add up really quickly”, says Potoczki. He continues to say that therefore, it was often beneficiary to set up distribution centers between producers and stores in which deliveries of different producers can be pooled in one transport unit to save costs. Potoczki deals with the question of how a company can use a network of distribution centers in the ideal way to make the most of it. In doing so, he has to take into account what seems to be an endless list of factors: Which locations are suitable geographically and where are the producers located? Do such distribution centers have storage capacity or are they merely designed as a hub for the goods? Which products are in high demand and which goods are rarely sold? “The diverse network and interdependency of these questions results in a very complex planning task which is difficult to solve but also holds much potential for cost-saving effects”, says Potoczki. For his doctoral thesis, Potoczki is currently developing a mathematical model which helps to design the best possible distribution network while Daniel Schubert focuses on the processes within a distribution center.

He seeks to optimize interplay of the different processes for the provision of goods and their delivery to customers and stores. “My aim is to achieve integrated planning of warehouse management and delivery. In simple terms, it is only possible to deliver the goods on time if they were picked and loaded in due time”, says Schubert. Up to now, he has already investigated the corresponding processes at a large retailer in Germany and a large food retailer. “In these cases, an automated and integrated approach to order picking and delivery shows on the one hand that there is potential to save up to more than 10 per cent of the costs, on the other hand it is also possible to avoid more than 50 per cent of delays in delivery.” To this end, Schubert currently develops different algorithms which offer insight into the respectively active processes and show possibilities for optimization.

 Markus Frank and Stefan Voigt took a closer look at an everyday phenomenon: As a general rule, the mailman always tries to deliver eagerly awaited parcels when the recipient is not home. This causes frustration for the customer and leads to higher costs for the delivery service provider. The researchers now want to use customers’ personal data, such as the history of past successful deliveries to be able to derive the probability of a successful delivery. Like this, the service provider could know when customers are most likely to be home or in which order they should best deliver the parcels. This could reduce costs which are incurred by repetitive driving distances of failed deliveries. “You can calculate that as few as four customers already result in 24 different possibilities for arranging deliveries on a tour; with ten customers we already have more than three million possibilities. This clearly shows that we need the help of specialized software to be able to develop a concept”, say Frank and Voigt, who currently work on a corresponding program.