Accurately pricing products to generate maximum revenue is no easy task. And that pricing challenge is further complicated when those products have a limited inventory, a finite selling season, multiple variations, and unknown consumer demand.

To solve this riddle, Yiwei Chen and his colleague propose a novel, nonparametric learning algorithm, called the online inverse batch gradient descent (IGD) algorithm, which follows an iterative process of estimating reference prices to market shares at batches of time intervals to identify optimal pricing.

Chen’s algorithm is recommended to firms, as it surpasses the accuracy of existing methods when pricing products in a market with unknown demand.