This paper describes the application of reinforcement learning (RL) to multi-product inventory management in supply chains. The problem description and solution are both adapted from a real-world business solution. The novelty of this problem with respect to supply chain literature is (i) we consider concurrent inventory management of a large number (50 to 1000) of products with shared capacity, (ii) we consider a multi-node supply chain consisting of a warehouse which supplies three stores, (iii) the warehouse, stores, and transportation from warehouse to stores have finite capacities, (iv) warehouse and store replenishment happen at different time scales and with realistic time lags, and (v) demand for products at the stores is stochastic. We describe a novel formulation in a multi-agent (hierarchical) reinforcement learning framework that can be used for parallelised decision-making, and use the advantage actor critic (A2C) algorithm with quantised action spaces to solve the problem. Experiments show that the proposed approach is able to handle a multi-objective reward comprised of maximising product sales and minimising wastage of perishable products.
Almost half of UK grocery retail directors say replenishment is still driven by gut feel, according to research by Blue Yonder, which supplies predictive applications of retail. It said that interviews with 750 grocery managers and directors in the USA, UK, Germany and France, showed that despite a rise in accurate machine learning algorithms for automated replenishment and demand planning, 46 per cent of surveyed directors in the UK say replenishment is still an entirely manual process. Some 85 per cent of respondents identified automation as a key tool for making the fast decisions needed to meet customer demand. The research also identified that 31 per cent of directors in the UK feel there are now too many decisions to be made manually, with the same number stating that gut feel is slowing them down. The research also found that 62 per cent of UK directors say they have invested in replenishment optimisation in the last two years; 31 per cent say they will be investing further in replenishment optimisation in the next two years.
WBOC-TV reports the U.S. Army Corps of Engineers' website now lists June 1 as the start date for the first stage of the project, at Bethany Beach. In February, Army Corps of Engineers spokesman Steve Rochette said a limited number of dredges nationwide had delayed the schedule for work in Bethany Beach, South Bethany Beach, and Fenwick Island.
New research has shown that grocery retailers are struggling to optimise stock replenishment processes, with almost half saying that their decisions are still based on'gut feeling'. Retail applications provider Blue Yonder surveyed 750 grocery managers and directors in the US, UK, Germany and France. It found that, in spite of a rise in accurate algorithms for automated replenishment and demand planning, 46% of surveyed directors in the UK say that replenishment is still an entirely manual process and the same amount saying that it was fully automated. A further 30% believed that instinct-based decision making was slowing them down. Of the four countries involved in Blue Yonder's survey, Germany had the highest proportion of respondents using manual or partially automated systems, with just one-third of managers who had fully automated their stock replenishment processes.