Trend

Calculating a fractional rate of change to be used in forecasting

Diagram

Information

This information is part of the Business Simulation Library (BSL). Please support this work and ► donate.

The Trend component represents forecasting in behavioral models and has been proposed by Sterman [3, pp. 634 - 638]. The basic idea is rather intuitive: Forecasters perceive a present condition of some input as a smoothed input (ppc), while they also perceive the historic or reference condition (rc) as a smooth of their current perception.

Dividing the diffence between the current condition and the reference condition by the reference condition will determine the fractional increase. A further division by the time horizon for the reference condition will then turn this into a fractional rate. The thus calculated rate of change (i.e., the indicated trend) may then again be smoothed, if smoothTrend = true.

See also

TrendBasedForecast, Smooth

Parameters (7)

initialTrend

Value: 0

Type: Rate (1/s)

Description: Initial trend

tppc

Value: 1

Type: Time (s)

Description: Averaging time for perception of current quantity (i.e., the present condition)

thrc

Value: 1

Type: Time (s)

Description: Time horizon for reference condition

tpt

Value: 1

Type: Time (s)

Description: Time to perceive the trend (i.e., there is gradual adaptation to a new trend)

smoothTrend

Value: false

Type: Boolean

Description: = true, if the calculated trend is to be smoothed

reportPPC

Value: false

Type: Boolean

Description: = true, if the perceived present conditon is to be reported

init

Value: modelSettings.init

Type: InitializationOptions

Description: Provide InitializationOptions (Free, FixedValue, SteadyState)

Connectors (3)

y

Type: RealOutput

Description: Information output signal

u

Type: RealInput

Description: Current value as basis for trend formulation

y_ppc

Type: RealOutput

Description: Report the perceived present condition (optional)

Components (11)

modelSettings

Type: ModelSettings

Description: Setting important global variables within a model

actualTrend

Type: PassThrough

Description: Output is identical to input

rc

Type: Smooth

Description: Reference condition for calculating the fractional trend

ppc

Type: Smooth

Description: Perceived present condition

parTPPC

Type: ConstantConverterTime

Description: Time to perceive present condition

parTHRC

Type: ConstantConverterTime

Description: Time horizon for the reference condition (duration of trend formulation)

parTPT

Type: ConstantConverterTime

Description: Time to perceive the trend (i.e., adjustment time)

perceivedTrend

Type: Smooth

Description: Perceived growth rate

indicatedTrend

Type: Division_Guarded

Description: Indicated fractional growth rate (trend)

delta

Type: Gap

Description: Difference between perceived present condition and reference condition

rcTimesTHRC

Type: Product_2

Description: Referenc condition times time horizon for reference condition

Used in Components (1)

TrendBasedForecast

BusinessSimulation.MoleculesOfStructure.InformationProcessing

Trend-based forecast for an exponentially growing quantity