Neural controller for business yield management

Abstract

In the business and operational research field there is a class of perishable inventory control problems called ‘yield management’. Examples are floating pricing strategies in air ticket reservation and hotel room booking. Due to the complex nature of yield management, there are few analytical models available for practical application. This paper presents a neural network approach to solving yield management problems. Using modified back propagation neural networks, a threshold band in the high-dimensional yield management space is generated based on historical data and/or management expertise. When the actual inventory level is outside the threshold band, prices should be adjusted to lead the inventory level back to the threshold band. The interval of the threshold band indicates the stability of the business system.