Artificial Intelligence (AI) – for example algorithms that create information to support decisions – can help decision makers to create business value and work more sustainably. The real value of AI lies in the creation of useful applications in practice. The ‘AI in the Floriculture Chain’ (iFlow) project explored the potential value of AI and developed AI applications with value for growers, auctioneers, transporters, and supply chain managers of flower products. The iFlow project – which is a collaboration of Royal FloraHolland, Zentoo, Wageningen University, and the Erasmus Center for Data Analytics at Rotterdam School of Management, Erasmus University (RSM) – led to key business and academic insights, which are useful for both the flower industry and the AI academic community.
Improvements on channel strategy and pricing
Growers decide about their multichannel strategy, for example when and how many products to sell via pre-sales, auction or direct sales. Zentoo is a grower association with 14 growers producing high-quality chrysanthemums. They sell flowers via the Royal FloraHolland auction channel and directly via their Zentoo web shop, which is also connected to Royal Flora Holland’s digital platform, Floriday. A model predicting next-day auction prices was developed to better align the web shop prices to the predicted auction prices of Zentoo’s chrysanthemums. The results indicate a positive price effect when Zentoo’s web shop manager uses the prediction model for price decisions.
For growers it is an important decision to put their products in pre-sales or not. Data were analysed to see the effects of pre-sales on volume sold. The results showed that even putting products in pre-sales without a pre-sales transaction, do have a positive impact on the auction sales of products.
Buyers’ behaviour was investigated. The empirical analysis unveiled a number of conditions when buyers may use both channels; when they have a large portfolio size, when there is friction in the pre-sales channel, or when they strategically complement both channels. There is evidence that if buyers use the pre-sales for the day and also join the auctions, they are more likely to opt for offline auction and also tend to enter later auctions. However, unlike these decisions that can be pre-planned at day level, the researchers did not find evidence relating to pre-sales choice with the time of entering the auction lot.
An econometric model and a simulation model were developed to advise Royal FloraHolland’s auctioneers on optimal minimal transaction amount and disclosure of the winning bidder. The detailed analysis of historical transactions which indicates the effect of different scenarios is a break-through both academically and practically.
Experiments showed that the non-disclosure of the winning bidder (identification) leads to higher and more stable prices. After detailed analysis Royal Flora Holland decided not to disclose the winning bidder anymore for all auction clocks.
Improvements in transport and logistic processes
The agro-food sector is a very data-rich domain. The challenge is how this data can be used to improve the logistics and supply chain processes and create additional benefit for the stakeholders. At the same time, although the value of visibility and information sharing has been discussed for a long period of time, some well-defined methods to design the visibility in the agro-food industry are still lacking. With these perspectives, four data analytics applications that support logistics decisions in the floriculture chain network, at both individual and supply-chain level, were investigated.
The first decision deals with the real-time workforce adjustment at the cross-docking facilities. Based on the historical time scan data, the historical and the real-time European Data Format, inbound volume can be predicted at a smaller, hourly, time horizon and could be used to decide what workforce is needed for the next days.
The second decision has to do with the design of storage and fulfillment services at the cross-docking. Royal FloraHolland would like to provide new services to their suppliers and customers, and an algorithm was developed and tested. This could lead to a new innovative service, anticipatory shipping, for example to ship flowers to customers before they place an order. This is a very counter-intuitive strategy: flowers are shipped before the customer orders them. Predictions – based on previous orders and other customer data – forecast when customers will buy specific flowers again and the flowers are already distributed to the customer before they order the flowers, which leads to more in-time delivery.
The third decision is about the delivery postponement in real-time process co-ordination. Collaboration and co-ordination among suppliers, the cross-docking facility, and customers is facilitated. Data of the farmer and greenhouse departure scan can help to predict the arrival time of the products at the cross-docking station and help to improve inbound scheduling. Cross-docking aims to distribute trolleys within one hour, but customers do not always have the required storage capacity or workers to handle the incoming trolleys. Cross-docking can postpone the delivery of trolleys when the customer boxes are too busy and can avoid long queue of trolleys. The developed simulation model showed that timely postponement signals from many customers to the cross-docking operator is crucial to gain more benefit for this type of co-ordination and could help customers to reduce total cross-docking workload.
The fourth decision focuses on strategic partner selection in horizontal collaboration. Historical time scan data provide details about arrival times of products from different suppliers at the cross-docking. A frequent pattern mining algorithm is developed with the data and determined patterns, for example a set of suppliers that arrive frequently at the same time at the Royal FloraHolland inbound docks. A substantial number of suppliers arrive at the same time, and that could potentially lead to a discussion among these suppliers to share truck capacity which helps them to lower transportation costs and carbon output.
The iFlow project, shows that the floriculture sector can be optimised and made more sustainable with the help of AI and AI applications.
To learn more, look into short and longer programmes in Technology and Operations for executives and managers.
Rotterdam School of Management, Erasmus University (RSM) is one of Europe’s top-ranked business schools. RSM provides ground-breaking research and education furthering excellence in all aspects of management and is based in the international port city of Rotterdam – a vital nexus of business, logistics and trade. RSM’s primary focus is on developing business leaders with international careers who can become a force for positive change by carrying their innovative mindset into a sustainable future. Our first-class range of bachelor, master, MBA, PhD and executive programmes encourage them to become to become critical, creative, caring and collaborative thinkers and doers. Study information and activities for future students, executives and alumni are also organised from the RSM office in Chengdu, China. www.rsm.nl
For more information about RSM or this article, please contact Danielle Baan, Media Officer for RSM, via +31 10 408 2028 or firstname.lastname@example.org.
Photocredit: Royal FloraHolland
Read more at: https://discovery.rsm.nl/articles/427-optimise-floriculture-and-make-it-more-sustainable-with-ai-algorithms/