Logistics platforms have experienced an unprecedented wave of robotization in recent years.Undaunted by the pandemic, this trend has been reinforced by changes in consumer behaviour, with an increase in e-commerce orders, particularly for food products (see the 2020 Sprint barometerProject, of which Savoye is a partner). It is motivated by the need to increase the overall performance of a platform, improve the quality of operations and gain execution responsiveness. This wave is worldwide, because it concerns Europe, North America, East Asia and the Middle East. This wave is global, because it is part of the strategy of Supply Chain organizations of large international groups as well as those of local players. And this wave is both multi-sectoral, impacting both B2C and B2B activities, and multi-business, structuring production and distribution sites, like those of logistics providers. This wave is marked by multiple parallel dynamics of innovations: technological innovations of course, but also increased intelligence of flow management software, or even the integration of Machine Learning contributions to facilitate activity forecasts.
Technological innovation: from automation to robotization
Omnichannel is a key factor in this evolution. The warehouse model dealing with only one channel is tending to disappear in favor of logistics platforms ensuring the preparation of orders from multiple sales channels. Each sales channel has its own logistics profile: number of lines per order and pieces per line, shipping time, etc. Each has its own volume and growth forecasts, and each has its appropriate order preparation process. Which means, when we talk about automation and robotization, that there is no single technology allowing all channels to be processed. And that there is not necessarily relevance, in the sense of economic interest, to automate all processes and all flows. In other words, the warehouse of tomorrow is not 100% robotic, but made up of multiple parallel processes, which can each be 100% manual, mechanized, automated or completely robotic. For a logistics decision-maker, choosing the appropriate technology for automating a process can quickly become a headache. The diversity of solutions offered has never been so important, and it is not easy to estimate their level of technical maturity and compare them with equivalent scope. This is where Savoye's added value lies, as architect of the warehouse flow plan from which we propose the solutions and technologies to be implemented. To do this, Savoye relies on a range of tools and components, either those we have available to us or via our partners.
The robotization of intralogistics operations is a further step taken in this global trend towards automation. First of all with the introduction of mobile robots, either operating in a secure enclosure and/or along a marked path (AGV), or completely autonomous and working with humans in the same environment (AMR). These robots can carry out missions of transporting loads, moving products in stock or orders to be prepared. Flexibility and scalability are its key assets. Then, with the robotization of retail picking operations, more particularly relevant in goods-to-person preparation processes (products to humans) where only the movements of picking up and putting down items are carried out manually. Although current technological maturity allows for the first relevant deployments of retail picking robots, steps remain to be taken before considering deployments on a larger scale. Savoye invests in these areas with expert and specialized partners, and is able to offer turnkey solutions to its customers. Savoye will also intervene on this subject during the Smart Supply Village of SprintProject during the next SITL, at Paris Porte de Versailles, from September 13th to 15th, 2021.
WES: the importance of algorithms
The implementation in the same diagram of differentiated preparation processes, either parallel or successive, requires adequate software management to manage the necessary processes of scheduling, prioritization or even consolidation of order flows. If a good level WMS will natively handle these functionalities in the case of completely manual processes, the use of a WES is essential to synchronize multiple automated flows. Not only does the WES communicate in real time with all the equipment, but it also acts constantly to optimize its use and prevent shortages or saturation of work to be carried out. A natural evolution of WCS, WES has a 360° vision of all flows and technologies used, with the objective of fluidification.
Proposing such a management solution is based on two major challenges. The first of these is the ability to offer such a solution on a robust technological platform, capable of absorbing extremely large volumes of data with instantaneous response times. The second challenge is the ability to propose the appropriate algorithms allowing the optimization of the system. This implies significant efforts in operational research applied directly to automated and robotic processes. These efforts are based on modeling and simulation work, representing a digital twin of a system, making it possible to define its optimal management. The resulting algorithms, directly injected into the WES, are levers for performance gains, either through better productivity of equipment, or by the overall reduction in preparation time for the same order portfolio.
Machine Learning: better predict to better anticipate
Major source of innovation for the Supply Chain and intralogistics in particular over the coming years: Machine Learning. Today the warehouse manager juggles between, on the one hand, the information available – the known orders to prepare, the human and material resources to respond to them, the stock status and the schedule of transporter appointments; on the other hand, the unknown information that arrives as it flows – the arrival of new orders to prepare, the actual execution time of operations, and of course the hazards!
The idea of using Machine Learning is to provide the warehouse manager with forecasts that draw on the past to better anticipate tomorrow. This field of application of Machine Learning, Forecasting, applies to many use cases inside the warehouse. First and foremost is Labor Management, with the ability to predict the number of operators needed per intralogistics process on a given day. Also Slotting, in order to anticipate the reallocation movements of references according to the evolution of their rotation class. Or even Predictive Maintenance, in order to optimize the availability rate of equipment.
Robotization: a necessary quest for meaning
Whether we talk about automation and robotization, the idea is to invest where it makes sense, with appropriate technologies. With the dual objective of adopting a solution that is economically relevant and easily appropriable by the teams. This requires an ergonomic and stress-free operating mode for operator stations, in order to build their loyalty more easily. Investing in a robotization project, whatever its scale, means becoming part of the history of the company: the solution must be durable and long-lasting in terms of growth prospects and future challenges.
Read all the “Expert Opinion” articles on the SprintProject blog
