AGIFORS--RM Study Group New York City, March 2000 Lawrence R. Weatherford, PhD University of Wyoming Unconstraining Methods.

Slides:



Advertisements
Similar presentations
Evaluating the Effects of Business Register Updates on Monthly Survey Estimates Daniel Lewis.
Advertisements

1 The Impact of Buy-Down on Sell Up, Unconstraining, and Spiral-Down Edward Kambour, Senior Scientist E. Andrew Boyd, SVP and Senior Scientist Joseph Tama,
Modeling Sell-up in PODS enhancements to existing sell-up algorithms, etc. Hopperstad March 00.
Chapter 16 Inferential Statistics
Presented to AGIFORS YM Study Group Bangkok, Thailand May 2001 Larry Weatherford University of Wyoming Dispersed Fares within a Fare Class: How Can We.
Confidential 1 DCPs in Forecasting Edward Kambour, Senior Scientist Roxy Cramer, Scientist.
Data Mining Methodology 1. Why have a Methodology  Don’t want to learn things that aren’t true May not represent any underlying reality ○ Spurious correlation.
Statistics Versus Parameters
Operations management Session 17: Introduction to Revenue Management and Decision Trees.
Statistics and Quantitative Analysis U4320
o & d Forecasting for O & D Control
PODS Update Large Network O-D Control Results
Improving Forecast Accuracy by Unconstraining Censored Demand Data Rick Zeni AGIFORS Reservations and Yield Management Study Group May, 2001.
Chapter 10 Section 2 Hypothesis Tests for a Population Mean
STAT 497 APPLIED TIME SERIES ANALYSIS
Topic 6: Introduction to Hypothesis Testing
Estimation from Samples Find a likely range of values for a population parameter (e.g. average, %) Find a likely range of values for a population parameter.
Cal State Northridge  320 Ainsworth Sampling Distributions and Hypothesis Testing.
1 Validation and Verification of Simulation Models.
CHAPTER 8 Estimating with Confidence
Lecture 17 Interaction Plots Simple Linear Regression (Chapter ) Homework 4 due Friday. JMP instructions for question are actually for.
“There are three types of lies: Lies, Damn Lies and Statistics” - Mark Twain.
1 EMSR vs. EMSU: Revenue or Utility? 2003 Agifors Yield Management Study Group Honolulu, Hawaii Larry Weatherford,PhD University of Wyoming.
Statistics for Managers Using Microsoft® Excel 7th Edition
Applied Business Forecasting and Planning
Airline On Time Performance Systems Design Project by Matthias Chan.
1 O&D Forecasting Issues, Challenges, and Forecasting Results John D. Salch PROS Revenue Management, Inc.
H oppersta d C onsultin g Market/Airline/Class (MAC) Revenue Management RM2003 Hopperstad May 03.
1 Sensitivity Studies of Network Optimization with Displacement Adjusted Virtual Nesting using PODS. Thomas Fiig, Revenue Management Development, Scandinavian.
CHAPTER 8 Estimating with Confidence
H OPPERSTA D C ONSULTIN G Market/Airline/Class (MAC) Revenue Management AGIFORS RM 2003 Hopperstad June 03.
ESTIMATES AND SAMPLE SIZES
Topics: Statistics & Experimental Design The Human Visual System Color Science Light Sources: Radiometry/Photometry Geometric Optics Tone-transfer Function.
Constrained Forecast Evaluation (CFE) Ronald P. Menich AGIFORS Res & YM 2-5 June 2003 HNL.
Choosing Sample Size for Knowledge Tracing Models DERRICK COETZEE.
Copyright © 2009 Pearson Education, Inc. Chapter 18 Sampling Distribution Models.
1 Chapter 18 Sampling Distribution Models. 2 Suppose we had a barrel of jelly beans … this barrel has 75% red jelly beans and 25% blue jelly beans.
RM Coordination and Bid Price Sharing in Airline Alliances: PODS Simulation Results Peter Belobaba Jeremy Darot Massachusetts Institute of Technology AGIFORS.
1 Arjan Westerhof 04 June 2016 Implementing an O&D System at KLM March 23, 2000 Arjan Westerhof Agifors yield management study group.
Chap 7-1 A Course In Business Statistics, 4th © 2006 Prentice-Hall, Inc. A Course In Business Statistics 4 th Edition Chapter 7 Estimating Population Values.
Statistical Inference for the Mean Objectives: (Chapter 9, DeCoursey) -To understand the terms: Null Hypothesis, Rejection Region, and Type I and II errors.
CHAPTER 9 Testing a Claim
Chap 7-1 A Course In Business Statistics, 4th © 2006 Prentice-Hall, Inc. A Course In Business Statistics 4 th Edition Chapter 7 Estimating Population Values.
Improving Revenue by System Integration and Cooperative Optimization Reservations & Yield Management Study Group Annual Meeting Berlin April 2002.
Lufthansa Update on LH´s Path of O&D Control AGIFORS, New York, March 2000 Werner Tauss.
Lufthansa Looking for Feedback Performance Measurement in Revenue Management Stefan Pölt Lufthansa German Airlines AGIFORS Reservations & Yield Management.
Sampling Fundamentals 2 Sampling Process Identify Target Population Select Sampling Procedure Determine Sampling Frame Determine Sample Size.
The Practice of Statistics, 5th Edition Starnes, Tabor, Yates, Moore Bedford Freeman Worth Publishers CHAPTER 7 Sampling Distributions 7.1 What Is A Sampling.
Chapter 8, continued.... III. Interpretation of Confidence Intervals Remember, we don’t know the population mean. We take a sample to estimate µ, then.
Sampling and Statistical Analysis for Decision Making A. A. Elimam College of Business San Francisco State University.
Introduction to Inference Sampling Distributions.
Math 3680 Lecture #15 Confidence Intervals. Review: Suppose that E(X) =  and SD(X) = . Recall the following two facts about the average of n observations.
1 Chapter 8 Interval Estimation. 2 Chapter Outline  Population Mean: Known  Population Mean: Unknown  Population Proportion.
Hypothesis test flow chart
The Practice of Statistics, 5th Edition Starnes, Tabor, Yates, Moore Bedford Freeman Worth Publishers CHAPTER 9 Testing a Claim 9.1 Significance Tests:
The inference and accuracy We learned how to estimate the probability that the percentage of some subjects in the sample would be in a given interval by.
Chapter 6: Descriptive Statistics. Learning Objectives Describe statistical measures used in descriptive statistics Compute measures of central tendency.
The accuracy of averages We learned how to make inference from the sample to the population: Counting the percentages. Here we begin to learn how to make.
The normal approximation for probability histograms.
Introduction Sample surveys involve chance error. Here we will study how to find the likely size of the chance error in a percentage, for simple random.
This represents the most probable value of the measured variable. The more readings you take, the more accurate result you will get.
Stats 242.3(02) Statistical Theory and Methodology.
Inference: Conclusion with Confidence
Sampling Distribution Models
Inference: Conclusion with Confidence
NATCOR Stochastic Modelling Course Inventory Control – Part 2
A Modeling Framework for Flight Schedule Planning
Estimating with PROBE II
Section 10.1: Confidence Intervals
Presentation transcript:

AGIFORS--RM Study Group New York City, March 2000 Lawrence R. Weatherford, PhD University of Wyoming Unconstraining Methods

Outline of Presentation I.Introduction II.Review of Common Unconstraining Methods III.Comparison of Performance IV. Conclusion

I. Introduction One of the major factors that affects forecast accuracy is the inability to observe the true (unconstrained) demand Improvements in forecast accuracy can translate into substantial revenue increases: Each 10% reduction in forecast error can be worth 2% to 4% in revenues on high demand flights (Lee, MIT Thesis)

RM forecasting requires a complete system that performs all of the following steps: Collection of historical data Cleaning of data (including outlier editing) Unconstraining of closed observations Estimation of forecast model from historical data Generate forecasts for each future flight Evaluate accuracy of forecasts/provide feedback to users

II. Review of Common Unconstraining Methods Going to look at the approach taken by 4 commonly used methods: A. Naive unconstraining (or detruncating) B. Pickup unconstraining C. Booking Curve unconstraining D. Projection unconstraining In unconstraining, we consider a class to be “closed” if at a booking period (reading day, DCP) the booking limit doesn’t allow for further bookings.

A. Naive unconstraining--only use unclosed observations (much more common than you might assume!) For example: Observation # Unclosed CC C  = ( )/3 = 20.33

B. Pickup unconstraining--increment closed observations by the higher of 1) average of the unclosed observations or 2) the actual value for the booking periods that are closed. Observation # Unclosed C8C C Unclosed3315 Avg Unclosed8.255 For example, using previous data:

Observation #2 Total for DCP’s 10-12: = Observation #3 Total for DCP’s 10-12: = 28  = ( )/5 = Note: just this simple method increases estimate of unconstrained demand from to 21.85

C. Booking Curve unconstraining--divide bookings-in-hand by long- run historical average ratio of bookings-in-hand to bookings at departure Using lots of historical data (not shown), suppose we determine that 75% of the bookings are received by DCP 10, and 85% by DCP 11. Observation # Unclosed CC C

Then for this example: Observation #2 Total for DCP’s 10-12: (50+ 7)/ = 26 Observation #3 Total for DCP’s 10-12: ( )/ = 36.5  = ( )/5 = 24.7

D. Projection unconstraining (statistically known as the EM method)-- uses much more complicated statistics that deal with “censored” observations Basic idea is to iterate at guessing the mean and standard deviation of the pickup from DCP First, you use the unclosed observations, then you find the conditional probabilities based on the constrained observations and re-estimate the , . This process continues until the values for ,  converge. Parameters: # of iterations, convergence limit critieria

Of course, all of these methods are further complicated by the following: 1) a given leg may be considered “open” and yet be closed to some of the O&D’s flowing over it due to some form of network control (e.g., bid price) 2) a booking class may be considered “open” because it was “open” on the 2 reading days, but could have actually been “closed” in between.

III. Comparison of Performance Intuitively, it makes sense that more statistically sound procedures like the “Projection” method should do a better job than the “Naïve” method at estimating the true unconstrained demand, but the question is how much better and is it worth the effort? Of course, one of the real problems in performing this analysis is that if one uses real airline data, we never know what the true unconstrained demand is and therefore are not able to accurately compare all 4 methods Leads us to use simulated data--randomly generated “true

unconstrained demand” and also randomly generated “booking limits” that determine whether or not we observe the true unconstrained demand or some constrained value. Then, we can make an honest evaluation of how much better one method does than another and how close it came to the true unconstrained demand (because we secretly know what that is).

A. Data Sets We’ll look at 5 different data sets (2 simulated data sets with 1000 observations each, 3 real data sets): 1. Simulated #1, unconstrained mean = 20, % unconstrained varies from 20% to 98% 2. Simulated #2, unconstrained mean = 4, % unconstrained varies from 20 to 98% 3. US Domestic, 14 fare classes, 6 departures, 90 days of data 4. European Continent, 10 fare classes, 14 departures,120 days 5. Carribean flight, 15 fare classes, 7 departures, 45 days

B. Results from Set #1 Avg improvement of EM over Naïve ranges from 7% to 47%, with an average increase of 21% across the 5 scenarios

C. Results from Set #2 Avg improvement of EM over Naïve ranges from 10% to 547%, with an average increase of 54% across the 5 scenarios

D. Results from Set #3 Avg improvement of EM over Pickup ranges from 18% to 1000%, with an average increase of 200% across the 14 fare classes

E. Results from Set #4 Avg improvement of EM over Pickup ranges from 10% to 200%, with an average increase of 36% across the 10 fare classes

F. Results from Set #5 Avg improvement of EM over Pickup ranges from 33% to 2000%, with an average increase of 97% across the 15 fare classes

G. Summary ·On average, using the EM method does much better than using either the Naïve or Pickup approaches (103% improvement in demand estimation across the 3 real data sets) · There is still the issue of what to do when all of your data for a given fare class is constrained, with all 0’s--EM can’t handle that

Questions? IV. Conclusion The type of unconstrainer you’re using can make a BIG difference.