The task of revenue management of an airline is to maximize the proceeds for the various products based market segmentation by price differentiation and product allocation.

As demand is subject to strong fluctuations for the individual products, this is a stochastic optimization problem. Numerous different modelings have been proposed over time for solution.

Remigius is a stochastic simulator. The original aim of the development was to evaluate the effects of exogenous factors on the optimization process. The simulation includes both the demand side (customers), as well as the supply side of the airlines.

Revenue Management

The task of revenue management results from the claims and properties of different customer groups. The more price-sensitive private customers want to fix their travel planning at a very early stage. The little price-sensitive business customers, however, usually know only a few days before departure if they are able to perceive an intended trip planning. A schedule-based airline relies on both clients. It offers customer groups associated products at different prices. To maximize the yield it is therefore a natural interest of the airline to reserve seats for the late booking business customers.

Thus, the revenue management needs to decide how many seats can be offered in advance under favorable conditions in order to still have enough seats for much paying business clients shortly before departure.


A revenue management simulator has to comprise the essential components of the real system. The accounting system will decide on the acceptance or rejection of booking requests of passengers and manages the inventory book of the simulated departures. The optimizer determined based on forecasts of future demand and the allocation of capacity requirements for blocks of seats for each customer class. Prognosis and fleet allocation provide the specifications. To illustrate the results of the simulation experiment, an output module is required with which the existing variety may be analyzed for key indicators at different levels. Special features of the simulation are the different ways to generate the demand and the possibility of a fleet assignment process.

Typical questions for simulations in the area of revenue management are:

  • Which optimization method dominates on equal terms?
  • How do the different assumptions of different optimization methods influence the result?

Another possible use is in the training of revenue managers.

Features of Regmigius

  • Demand:

    • Generation based on empirical data (volume and transaction history)
    • Generation based on stochastic processes (Poisson, Cox)
    • Upsell probabilities depending on customer groups and booking classes
    • Mapping of trends / seasonality in response to sales territories

  • Forecast

    • Basis exponential smoothing
    • Unconstraining by a pick-up model
    • Updating by a additiv pick-up model

  • Optimization

    • Alternatives: FCFS, EMSRb, EMSRd, MDP, DLP
    • Calculating an upper bound

  • Capacity assignment

    • Rolling fleet assignment with traffic-depending ground times

  • Net

    • Two calibrated networks

      • European netzwork with 730 flights and 1605 O&Ds
      • European netzwork with 180 flights and 880 O&Ds

    • Calibration of other networks possible

  • Output analysis:

    • Bookings, Quotas and forecast grouped by O&Ds
    • Bookings, Quotas and forecast grouped by booking classes
    • Bookings, Quotas and forecast grouped by individual routes
    • Bookings, Quotas and forecast grouped by traffic areas
    • Bookings, Quotas and forecast in the course of time
    • Bookings, Quotas and forecast  grouped by partial fleets


Sitemap  Contact  Data Privacy  Imprint
© TU Clausthal 2022