ProBP Deprecated Approach¶

In this example, ProBP is applied with daily re-optimization, using a forecast adjusted by linear interpolation between the forecast point and departure. This is provided as a technical demonstration, but for actual implementations and analysis of ProBP reoptimization, users are recommended to employ the timeframe-based adjustment algorithm instead.

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/11-probp-daily.yaml")

show_file("network/11-probp-daily.yaml")

include:
  - 10-probp.yaml

scenario: 3MKTproBP-Daily

rm_systems:
  rm_probp:
    processes:
      DAILY:
        - step_type: forecastadjustment
          algorithm: timeframe
          kind: path
        - step_type: probp
          name: optimization

simulation_controls:
  show_progress_bar: true

db:
  write_items:
    - leg_final
    - fare_final
    - demand_final
    - bookings
    - bucket
    - pathclass
    - leg
    - demand
    - leg_daily
    - pathclass_daily

snapshot_filters:
  - type: pro_bp
    title: ProBP Snapshot
    sample: [ 120, 290, 499 ]
    dcp: [ 63, 56, 21 ]
    flt_no: 111
    airline: AL1
    directory: snapshots/probp
  - type: forecast_adj
    title: Forecast Adj Snapshot
    sample: [ 120, 290, 499 ]
    flt_no: 111
    airline: AL1
    directory: snapshots/fcst-adj

In [4]:

cfg = pax.Config.from_yaml([
    "network/11-probp-daily.yaml",
])

cfg = pax.Config.from_yaml([ "network/11-probp-daily.yaml", ])

In [5]:

cfg.rm_systems['rm_probp'].processes['daily'][0].algorithm = "days_to_departure"

cfg.rm_systems['rm_probp'].processes['daily'][0].algorithm = "days_to_departure"

In [6]:

sim = pax.Simulation(cfg)

sim = pax.Simulation(cfg)

In [7]:

summary = sim.run()

summary = sim.run()

Task Completed after 81.80 seconds

In [8]:

summary.fig_carrier_revenues()

summary.fig_carrier_revenues()

Out[8]:

In [9]:

summary.fig_carrier_load_factors()

summary.fig_carrier_load_factors()

Out[9]:

In [10]:

summary.fig_fare_class_mix()

summary.fig_fare_class_mix()

Out[10]:

In [11]:

summary.fig_bookings_by_timeframe()

summary.fig_bookings_by_timeframe()

Out[11]:

In [12]:

summary.to_xlsx("outputs/3mkt-12.xlsx")

summary.to_xlsx("outputs/3mkt-12.xlsx")

Comparing against Targets¶

In [13]:

import targets

target = targets.load(12, cfg)

import targets target = targets.load(12, cfg)

In [14]:

from passengersim import contrast

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

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

In [15]:

comps.fig_demand_to_come("mean") | comps.fig_demand_to_come("std")

comps.fig_demand_to_come("mean") | comps.fig_demand_to_come("std")

Out[15]:

In [16]:

comps.fig_bookings_by_timeframe(by_carrier="AL1")

comps.fig_bookings_by_timeframe(by_carrier="AL1")

Out[16]:

In [17]:

comps.fig_carrier_revenues()

comps.fig_carrier_revenues()

Out[17]:

In [18]:

comps.fig_carrier_load_factors()

comps.fig_carrier_load_factors()

Out[18]:

In [19]:

comps.fig_fare_class_mix()

comps.fig_fare_class_mix()

Out[19]:

In [20]:

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

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

Out[20]:

Interpolation to departure is not advised because customer arrival rates are often very far from uniformly distributed, resulting in forecasts that exhibit a sawtooth pattern.

In [21]:

print(sim.path_names()[1])
comps.fig_path_forecasts(by_path_id=1, of=['mu', 'sigma', 'closed'])

print(sim.path_names()[1]) comps.fig_path_forecasts(by_path_id=1, of=['mu', 'sigma', 'closed'])

Path: BOS ORD (AL1:101 BOS-ORD)

Out[21]:

In [22]:

print(sim.path_names()[5])
comps.fig_path_forecasts(by_path_id=5, of=['mu', 'sigma', 'closed'])

print(sim.path_names()[5]) comps.fig_path_forecasts(by_path_id=5, of=['mu', 'sigma', 'closed'])

Path: ORD LAX (AL1:111 ORD-LAX)

Out[22]:

In [23]:

print(sim.path_names()[9])
comps.fig_path_forecasts(by_path_id=9, of=['mu', 'sigma', 'closed'])

print(sim.path_names()[9]) comps.fig_path_forecasts(by_path_id=9, of=['mu', 'sigma', 'closed'])

Path: BOS LAX (AL1:101 BOS-ORD, AL1:111 ORD-LAX)

Out[23]:

In [24]:

comps.fig_bid_price_history(by_carrier="AL1", cap="some")

comps.fig_bid_price_history(by_carrier="AL1", cap="some")

Out[24]:

In [25]:

comps.fig_bid_price_history(by_carrier="AL1", cap="some", show_stdev=1)

comps.fig_bid_price_history(by_carrier="AL1", cap="some", show_stdev=1)

Out[25]:

In [26]:

from passengersim.extra.forecast_viz import fig_forecasts_and_bid_prices

from passengersim.extra.forecast_viz import fig_forecasts_and_bid_prices

In [27]:

fig_forecasts_and_bid_prices(
    sim,
    trial = 0,
    rrd = 63,
    flt_no = 111,
)

fig_forecasts_and_bid_prices( sim, trial = 0, rrd = 63, flt_no = 111, )

Out[27]:

In [28]:

fig_forecasts_and_bid_prices(
    sim,
    trial = 0,
    rrd = 63,
    flt_no = 101,
)

fig_forecasts_and_bid_prices( sim, trial = 0, rrd = 63, flt_no = 101, )

Out[28]: