Q Forecasting¶

In [1]:

from textwrap import fill

import numpy as np
import pandas as pd

import passengersim as pax

pax.versions()

from textwrap import fill import numpy as np import pandas as pd import passengersim as pax pax.versions()

passengersim 0.51.dev12+gb8c884e
passengersim.core 0.51.dev2+g02d3ce20

In [2]:

cfg = pax.Config.from_yaml(pax.demo_network("3MKT"))

cfg = pax.Config.from_yaml(pax.demo_network("3MKT"))

In this example, we will work with a completely fenceless marketplace. To convert the typical 3MKT network into a restriction-free network, we can use a tool to strip the restrictions.

In [3]:

from passengersim.config.manipulate import strip_all_restrictions

cfg = strip_all_restrictions(cfg)

from passengersim.config.manipulate import strip_all_restrictions cfg = strip_all_restrictions(cfg)

In [4]:

cfg.carriers.AL1.rm_system = "Q"
cfg.carriers.AL2.rm_system = "C"

cfg.carriers.AL1.rm_system = "Q" cfg.carriers.AL2.rm_system = "C"

We need to revalidate the congfiguration to be able to view or modify the RM system definitions prior to initializing the simulation.

In [5]:

cfg = cfg.model_revalidate()

cfg = cfg.model_revalidate()

In [6]:

print(fill(cfg.rm_systems.Q.description))

print(fill(cfg.rm_systems.Q.description))

Path-based additive pickup PODS-style Q-hybrid forecast (with no fare
restrictions this becomes Q-only), EM detruncation of priceable and
yieldable demand (separately), ProBP optimization.

In [7]:

print(fill(cfg.rm_systems.C.description))

print(fill(cfg.rm_systems.C.description))

Path-based additive pickup conditional forecast, EM detruncation of
yieldable demand, ProBP optimization.

In [8]:

cfg.simulation_controls.num_trials = 1
cfg.simulation_controls.num_samples = 600

cfg.simulation_controls.num_trials = 1 cfg.simulation_controls.num_samples = 600

In [9]:

sim = pax.Simulation(cfg)
summary = sim.run()

sim = pax.Simulation(cfg) summary = sim.run()

Task Completed after 4.83 seconds

In [10]:

summary.fig_carrier_revenues()

summary.fig_carrier_revenues()

Out[10]:

In [11]:

summary.fig_carrier_load_factors()

summary.fig_carrier_load_factors()

Out[11]:

In [12]:

summary.fig_carrier_rasm()

summary.fig_carrier_rasm()

Out[12]:

In [13]:

summary.fig_fare_class_mix()

summary.fig_fare_class_mix()

Out[13]:

In [14]:

summary.fig_carrier_head_to_head_revenue("AL1", "AL2")

summary.fig_carrier_head_to_head_revenue("AL1", "AL2")

Out[14]:

Without Priceable Detruncation¶

We can remove the detruncation of priceable demand from the Q system, and re-evaluate to see the difference.

In [15]:

cfg2 = cfg.model_copy(deep=True)
del cfg2.rm_systems.Q.processes.dcp.q_untruncation

cfg2 = cfg.model_copy(deep=True) del cfg2.rm_systems.Q.processes.dcp.q_untruncation

In [16]:

summary2 = pax.Simulation(cfg2).run()

summary2 = pax.Simulation(cfg2).run()

Task Completed after 5.60 seconds

In [17]:

summary2.fig_carrier_revenues()

summary2.fig_carrier_revenues()

Out[17]:

In [18]:

summary2.fig_carrier_load_factors()

summary2.fig_carrier_load_factors()

Out[18]:

In [19]:

summary2.fig_carrier_rasm()

summary2.fig_carrier_rasm()

Out[19]:

In [20]:

summary2.fig_fare_class_mix()

summary2.fig_fare_class_mix()

Out[20]:

In [21]:

summary2.fig_carrier_head_to_head_revenue("AL1", "AL2")

summary2.fig_carrier_head_to_head_revenue("AL1", "AL2")

Out[21]:

Applying Fare Adjustment¶

In [22]:

cfg3 = cfg.model_copy(deep=True)
cfg3.carriers.AL1.rm_system = "F"
cfg3.carriers.AL2.rm_system = "M"

cfg3 = cfg.model_copy(deep=True) cfg3.carriers.AL1.rm_system = "F" cfg3.carriers.AL2.rm_system = "M"

In [23]:

cfg3 = cfg3.model_revalidate()

cfg3 = cfg3.model_revalidate()

In [24]:

print(fill(cfg3.rm_systems.F.description))

print(fill(cfg3.rm_systems.F.description))

Path-based additive pickup PODS-style Q-hybrid forecast (with no fare
restrictions this becomes Q-only), EM detruncation of priceable and
yieldable demand (separately), MR fare adjustment with 25% scaling,
ProBP optimization.

In [25]:

print(fill(cfg3.rm_systems.M.description))

print(fill(cfg3.rm_systems.M.description))

Path-based additive pickup conditional forecast, EM detruncation of
yieldable demand, MR fare adjustment with 25% scaling, ProBP
optimization.

In [26]:

sim3 = pax.Simulation(cfg3)
summary3 = sim3.run()

sim3 = pax.Simulation(cfg3) summary3 = sim3.run()

Task Completed after 4.86 seconds

In [27]:

summary3.fig_carrier_revenues()

summary3.fig_carrier_revenues()

Out[27]:

In [28]:

summary3.fig_carrier_load_factors()

summary3.fig_carrier_load_factors()

Out[28]:

In [29]:

summary3.fig_carrier_rasm()

summary3.fig_carrier_rasm()

Out[29]:

In [30]:

summary3.fig_fare_class_mix()

summary3.fig_fare_class_mix()

Out[30]:

In [31]:

summary3.fig_carrier_head_to_head_revenue("AL1", "AL2")

summary3.fig_carrier_head_to_head_revenue("AL1", "AL2")

Out[31]: