
LPP and PSI Polska have implemented an innovative solution for optimizing order picking paths using artificial intelligence. The machine learning mechanism will be used in the Polish clothing company’s warehouse handling online sales. The new solution will shorten the distance required for order picking, increase facility efficiency, optimize the use of warehouse resources, and accelerate e-commerce shipping operations.
LPP, the owner of the Reserved, Cropp, House, Mohito, and Sinsay brands, which handles approximately 11 million orders annually, is one of the leaders in the e-commerce market in Poland and Europe. Online sales are growing, and the coronavirus epidemic has further influenced their momentum. E-commerce operations are becoming an increasingly significant challenge. Last year, the share of online sales in LPP’s business amounted to 12%.
– At the moment, we are observing a fourfold increase in customer interest in online shopping. This determines the need to optimize logistics and IT solutions aimed at meeting new customer expectations without compromising the quality and, above all, the speed of processing online orders – says Jacek Kujawa, Vice President of the Management Board of LPP. – That is why we decided, together with our long-term business partner – PSI Polska, to implement a solution based on artificial intelligence algorithms, which will allow us to significantly increase the efficiency of online order processing – he adds.
– The primary task of the implemented algorithm is to effectively solve the so-called Traveling Salesperson Problem (TSP). It consists of determining the shortest route connecting several points on a map. In the case of a warehouse, the system must determine the optimal route for dozens of picking locations. What seems simple at first glance is, in reality, a significant problem for mathematicians that has existed for years – says Jerzy Danisz, Head of PSIwms System Development at PSI.
LPP and PSI Polska decided to develop a solution using artificial intelligence algorithms. The use of a machine learning mechanism, specifically a neural network based on
CNN (Convolutional Neural Network) units, allows for generating lists of products needed to complete an order in real-time, determining the shortest path, and proposing optimal resource utilization, including equipment, carts, and warehouse staff time. This is possible, among other things, thanks to the application of a so-called genetic algorithm, which, through the application of evolutionary operators such as crossover and mutation, enables the creation of an optimal picking list.
– Thanks to the implemented solution, as the first tests show, the length of picking paths has been reduced by 30%. This also translates into a significant increase in the efficiency of the picking process itself in the warehouse by as much as 12.3% – explains Sylwester Dmytriwski, E-commerce Fulfillment General Manager at LPP. – The algorithm not only allows for shortening the distance needed to pick an order. Thanks to the artificial intelligence module, the mechanism “learns” based on the information it possesses. It absorbs data regarding the warehouse, orders, and available personnel. Based on this, it groups products, facilitating the entire process.
Constant cooperation with other modules of the resource management system enables real-time information updates and adjustment of the generated list. The picking path optimizer is part of the Warehouse Intelligence project. Its concept, based on the use of artificial intelligence, was appreciated by a team of experts from the National Centre for Research and Development in Warsaw and recommended for project funding.
– Good cooperation and the use of agile implementation methodologies allowed PSI and LPP to efficiently introduce an innovative solution that brought satisfactory results already in the first phase of the project. We have been working with LPP for years. During this time, the PSIwms system has allowed us to handle more than a tenfold increase in volume at the LPP distribution center in Pruszcz Gdański. We are pleased that LPP also chose us to handle its e-commerce processes. The LPP team’s ability to quickly implement innovations and respond to changes is commendable and inspiring for us – concludes Arkadiusz Niemira, President of the Management Board of PSI Polska.
Source: LPP








