An advanced order batching approach for automated sequential auctions with forecasting and postponement

Abstract

In sequential auctions, all the sub-orders from a buyer need to be sorted and consolidated within a short time window for shipping. Buyer demands and sub-order arrival times are uncertain. The current auction order fulfillment is facing several challenges. Based on a re-engineered Industrial Internet-of-Things (IIoT)-enabled automation system, this paper introduces an order batching approach with forecasting and postponement. Such an approach generates batches considering time interval and buyer completion rate to minimise the total processing time of the auction orders and system response time. The buyer completion rate refers to the ratio of current cumulative and predicted purchase quantity. We use the forecasting method proposed by Kong et al. (2021) to estimate the purchasing quantity. Through a series of computational experiments using real-life data, the proposed order batching method achieves a shorter order processing time and system response time. Results show that the number of auction buyers poses no effect on the performance of the proposed approach. Key parameters of order postponement rule influence on performance.