Optimisation algorithms to minimize energy consumption in automated warehouse

 Savoye is a company that provides hardware and software solutions to equip warehouses and, in general, logistics processes. Thanks to its expansion into the US and Asian markets, it is now growing strongly.
The FLOWER project, coordinated by Savoye, involves the laboratories LISPEN (ENSAM Lille) and DISP (INSA Lyon). The objective is to integrate energy cost criteria into the design and algorithmic management of logistics platforms in order to achieve a significant reduction in CO2 emissions. Modern logistics platforms are organized around the concept of “GTP” (Goods-To-Person). In this organization, the preparation of retail orders is carried out by an operator in a dedicated station, constantly fed by stock bins to take references and order boxes to be made. The scheduling of appointments between the boxes and order boxes, at the right time, in the right order, and without ever disconnecting the GTP station, is ensured by high-speed automated storage/retrieval systems (ASRS) and supported by proprietary management algorithms. The current design of ASRS and tests over their management algorithms is based on commercial simulators, used to optimize management rules to maximize productivity. In this way, the ASRS are designed to support a flow of orders during peak periods, but cannot adapt to the downward variation of an order book or take into account the environmental impact of operations. The climate crisis makes it imperative to take into account the cost of energy, in order to achieve a sustainable management of ASRS in real time.
This view is shared by an increasing number of Savoye customers, who are starting to consider the energy cost of ASRS systems into their decision criteria when calling for projects. The objective of this thesis is to formulate algorithms capable of ensuring order preparation while reducing the energy consumption of ASRS. In this context, the PhD student will work with the R&D department, which is attached to the group’s general direction and expert in AI, operational research and mathematical optimization. Two approaches are considered to meet the objective. The first is to reduce the speed of moving equipment, the second is to reduce the distance they travel. We will first study the algorithms that influence speed in order to achieve significant energy reductions while keeping unchanged most of the decision algorithms. Then, we'll study distance-based algorithms to address the following decision problems :
If an item required by a sampling station is present in several locations in the ASRS, from which location should its bin be called to minimize the overall system's energy consumption?
Storage Location Assignment Problem (SLAP): in which physical location the bin should be stocked after the picking operation ?
This last question is not only about looking for the best way to organize ASRS stock for more than 100,000 references in order to reduce energy consumption, but also to examine the impact of this storage mission when systems return to maximum performance. Strong attention will be paid to the applicability of algorithms developed on real full-size instances.