EM Untruncation¶

In this example, we include product and advance purchase (AP) restrictions in the simulation, and have each airline use the leg-based EMSR-B algorithm to manage revenue. Additionally, the EM algorithm is employed to detruncate censored demand. Demand is simulated at normal levels.

In [1]:

import passengersim as pax
pax.versions()

import passengersim as pax pax.versions()

passengersim 0.18.1
passengersim.core 0.18.1

In [2]:

from passengersim.utils.codeview import show_file

from passengersim.utils.codeview import show_file

In [3]:

show_file("network/08-untrunc-em.yaml")

show_file("network/08-untrunc-em.yaml")

include:
  - 01-base.yaml
  - 02-buyup.yaml
  - 03-ap.yaml

simulation_controls:
  demand_multiplier: 1.0

rm_systems:
  rm_test1:
    availability_control: leg
    processes:
      DCP:
      - step_type: untruncation
        name: untruncation
        algorithm: em
        kind: leg
      - step_type: forecast
        name: forecast
        algorithm: additive_pickup
        alpha: 0.15
        kind: leg
      - step_type: emsr
        name: optimization
        algorithm: emsrb
        kind: leg

airlines:
  - name: AL1
    rm_system: rm_test1
  - name: AL2
    rm_system: rm_test1

In [4]:

cfg = pax.Config.from_yaml([
    "network/08-untrunc-em.yaml",
])

cfg = pax.Config.from_yaml([ "network/08-untrunc-em.yaml", ])

In [5]:

cfg.simulation_controls.num_trials = 6

cfg.simulation_controls.num_trials = 6

In [6]:

cfg.outputs.reports.add("local_and_flow_yields")

cfg.outputs.reports.add("local_and_flow_yields")

In [7]:

sim = pax.Simulation(cfg)

sim = pax.Simulation(cfg)

In [8]:

summary = sim.run()

summary = sim.run()

Task Completed after 13.59 seconds

In [9]:

summary.fig_carrier_revenues()

summary.fig_carrier_revenues()

Out[9]:

In [10]:

summary.fig_carrier_load_factors()

summary.fig_carrier_load_factors()

Out[10]:

In [11]:

summary.fig_fare_class_mix()

summary.fig_fare_class_mix()

Out[11]:

In [12]:

summary.fig_bookings_by_timeframe()

summary.fig_bookings_by_timeframe()

Out[12]:

In [13]:

summary.local_and_flow_yields

summary.local_and_flow_yields

Out[13]:

	flt_no	carrier	orig	dest	capacity	distance	local_yield	local_fraction	flow_yield
0	101	AL1	BOS	ORD	100	863.753282	0.228168	0.419748	0.160276
1	102	AL1	BOS	ORD	100	863.753282	0.227869	0.416923	0.161073
2	111	AL1	ORD	LAX	120	1739.799337	0.175814	0.545423	0.160276
3	112	AL1	ORD	LAX	120	1739.799337	0.175983	0.546765	0.161073
4	201	AL2	BOS	ORD	100	863.753282	0.228095	0.419208	0.161365
5	202	AL2	BOS	ORD	100	863.753282	0.228021	0.420197	0.161306
6	211	AL2	ORD	LAX	120	1739.799337	0.176204	0.545464	0.161365
7	212	AL2	ORD	LAX	120	1739.799337	0.176278	0.547227	0.161306

In [14]:

summary.to_xlsx("outputs/3mkt-08.xlsx")

summary.to_xlsx("outputs/3mkt-08.xlsx")

Comparing against Targets¶

In [15]:

import targets
target = targets.load(8, cfg)

import targets target = targets.load(8, cfg)

In [16]:

from passengersim import contrast

from passengersim import contrast

In [17]:

comps = contrast.Contrast({
    "simulation": summary,
    "target": target,
})

comps = contrast.Contrast({ "simulation": summary, "target": target, })

In [18]:

comps.fig_demand_to_come("mean")

comps.fig_demand_to_come("mean")

Out[18]:

In [19]:

comps.fig_demand_to_come("std")

comps.fig_demand_to_come("std")

Out[19]:

In [20]:

comps.apply(
    lambda s: (
        s.aggregate_demand_history(by_segment=True).unstack("segment").cov()
    )
)

comps.apply( lambda s: ( s.aggregate_demand_history(by_segment=True).unstack("segment").cov() ) )

Out[20]:

	segment	business	leisure
source	segment
simulation	business	3263.099012	1976.758878
	leisure	1976.758878	5291.913397
target	business	3249.598992	1955.419868
	leisure	1955.419868	5359.555328

In [21]:

contrast.fig_bookings_by_timeframe(comps, by_carrier="AL1",  by_class=True)

contrast.fig_bookings_by_timeframe(comps, by_carrier="AL1", by_class=True)

Out[21]:

In [22]:

comps.fig_carrier_revenues()

comps.fig_carrier_revenues()

Out[22]:

In [23]:

comps.fig_carrier_load_factors()

comps.fig_carrier_load_factors()

Out[23]:

In [24]:

comps.fig_fare_class_mix()

comps.fig_fare_class_mix()

Out[24]:

In [25]:

comps.fig_bookings_by_timeframe(by_carrier="AL1", by_class=True)

comps.fig_bookings_by_timeframe(by_carrier="AL1", by_class=True)

Out[25]:

In [26]:

comps.fig_leg_forecasts(by_flt_no=101, of=["mu", "sigma"])

comps.fig_leg_forecasts(by_flt_no=101, of=["mu", "sigma"])

Out[26]:

In [27]:

comps.fig_leg_forecasts(by_flt_no=101, of=["mu", "sigma"])

comps.fig_leg_forecasts(by_flt_no=101, of=["mu", "sigma"])

Out[27]:

In [28]:

comps.fig_leg_forecasts(by_flt_no=111, of=["mu", "sigma"])

comps.fig_leg_forecasts(by_flt_no=111, of=["mu", "sigma"])

Out[28]: