WEBVTT
00:00:09.500 --> 00:00:11.960
To illustrate how linear
optimization works in revenue
00:00:11.960 --> 00:00:14.290
management, let us consider
a simple example --
00:00:14.290 --> 00:00:17.620
a flight from New
York to Los Angeles.
00:00:17.620 --> 00:00:20.980
In this flight, there are
two types of economy fares,
00:00:20.980 --> 00:00:25.930
Early Bird fares that cost
$238, and Last Minute fares
00:00:25.930 --> 00:00:29.170
that cost $617.
00:00:29.170 --> 00:00:32.320
In this flight, a
Boeing 757 is used
00:00:32.320 --> 00:00:36.720
that has 166 economy seats.
00:00:36.720 --> 00:00:39.300
Demand for these prices
has been forecasted
00:00:39.300 --> 00:00:42.390
using analytics tools,
looking at historical data
00:00:42.390 --> 00:00:44.220
and incorporating
models like time
00:00:44.220 --> 00:00:46.520
series or linear regression.
00:00:46.520 --> 00:00:49.920
Clearly, forecasts have
errors, and therefore, we
00:00:49.920 --> 00:00:53.200
need to assess the
sensitivity of our decisions
00:00:53.200 --> 00:00:54.980
to these errors.
00:00:54.980 --> 00:00:57.060
To illustrate the use
of linear optimization,
00:00:57.060 --> 00:00:59.470
we assume that demand has
already been forecasted.
00:01:02.220 --> 00:01:05.650
We'll illustrate how our
decisions on how many discount
00:01:05.650 --> 00:01:11.860
seats to sell vary as the
demand forecasts vary.
00:01:11.860 --> 00:01:14.340
If the demand for
regular seats is
00:01:14.340 --> 00:01:19.500
50, and for discounted
fares is 150,
00:01:19.500 --> 00:01:24.250
and the capacity is 166 seats,
then the optimal allocation
00:01:24.250 --> 00:01:28.800
is going to be to sell
the 50 seats to satisfy
00:01:28.800 --> 00:01:33.610
the regular demand, and then
we allocate the remaining 116
00:01:33.610 --> 00:01:37.770
seats to the
discounted fare class.
00:01:37.770 --> 00:01:40.700
If the regular demand
increases to 100 seats,
00:01:40.700 --> 00:01:46.080
then we allocate these 100 seats
to these customers, and only
00:01:46.080 --> 00:01:49.910
66 seats to discounted
fare customers.
00:01:49.910 --> 00:01:53.710
Finally, if the regular
demand increases to 200,
00:01:53.710 --> 00:01:58.130
then we allocate all of
our capacity, 166 seats,
00:01:58.130 --> 00:02:00.260
to these customers.
00:02:00.260 --> 00:02:03.200
While this seems
simple, what happens
00:02:03.200 --> 00:02:09.830
if we have 100 flights with
connections in tens of fares?
00:02:09.830 --> 00:02:13.480
We'll next see how to formulate
the problem mathematically
00:02:13.480 --> 00:02:17.920
and solve it in a systematic
way, using linear optimization.