Understanding Air Transportation Market Dynamics using a Search Algorithm for Calibrating Travel Demand and Price

Abstract

This paper presents a search algorithm based framework to calibrate origin-destination (O-D) market specific airline ticket demands and prices for the Air Transportation System (ATS). This framework is used for calibrating an agent based model of the air ticket buy-sell process -Airline Evolutionary Simulation (Airline EVOS) -that has fidelity of detail that accounts for airline and consumer behaviors and the interdependencies they share between themselves and the NAS. More specificially, this algorithm simultaneous calibrates demand and airfares for each O-D market, to within specified threshold of a pre-specified target value. The proposed algorithm is illustrated with market data targets provided by the Transportation System Analysis Model (TSAM) and Airline Origin and Destination Survey (DB1B). Although we specify these models and datasources for this calibration exercise, the methods described in this paper are applicable to calibrating any low-level model of the ATS to some other demand forecast model-based data. We argue that using a calibration algorithm such as the one we present here to synchronize ATS models with specialized forecast demand models, is a powerful tool for establishing credible baseline conditions in experiments analyzing the effects of proposed policy changes to the ATS. Nomenclature α= price elasticity of demand λ = multiplier for airfare of non-stop itineraries A = advertised airfare B = basefare d = number of days to travel date l = load-factor (ratio of sold to total number of available seats)

An Integrated Framework for Modeling Air Carrier Behavior, Policy, and Impacts in the U.S. Air Transportation System

“…Necessary calculations are made for assigning individual agent characteristics and other behavior-driving factors. 2) Once the AIRLINE-EVOS system has been set up, a calibration loop-indicated by a dashed line-is started to match AIRLINE-EVOS simulated outcomes for airfare and demand to TSAM projections, as described in Kumar 6 . This calibration process implemented a search algorithm that iteratively adjusted on a market level, airfares and customer willingness-to-pay.…”

“…Thus the customer willingness to pay (WTP) attribute is integrally linked with the size of a market's demand. We then calibrate for each market, its WTP curve through a search algorithm described in Kumar 6 . In this way, we get close to accurately matching the level of market demand projected by TSAM scenarios through approximating customer decisions to not purchase a ticket if all of the offered airfares are more expensive than the maximum amount they are willing to pay.…”