| PowerMultiplicativeCooling {xegaPopulation} | R Documentation |
Power multiplicative cooling.
Description
This schedule decreases by the inverse proportion of
a power
of k. lF$Alpha() should be larger than 1.
Usage
PowerMultiplicativeCooling(k, lF)
Arguments
k |
Number of steps to discount. |
lF |
Local configuration. |
Details
Temperature is updated at the end of each generation
in the main loop of the genetic algorithm.
For lF$CoolingPower()==1 and
lF$CoolingPower()==2 this results in the
the linear and quadratic multiplicative cooling schemes
in A Comparison of Cooling Schedules for Simulated Annealing.
lF$Temp0() is the starting temperature.
lF$Alpha() is a scaling factor.
lF$CoolingPower() is an exponential factor.
Value
Temperature at time k.
References
The-Crankshaft Publishing (2023) A Comparison of Cooling Schedules for Simulated Annealing. <https://what-when-how.com/artificial-intelligence/a-comparison-of-cooling-schedules-for-simulated-annealing-artificial-intelligence/>
See Also
Other Cooling:
ExponentialAdditiveCooling(),
ExponentialMultiplicativeCooling(),
LogarithmicMultiplicativeCooling(),
PowerAdditiveCooling(),
TrigonometricAdditiveCooling()
Examples
parm<-function(x){function() {return(x)}}
lF<-list(Temp0=parm(100), Alpha=parm(1.01), CoolingPower=parm(2))
PowerMultiplicativeCooling(0, lF)
PowerMultiplicativeCooling(2, lF)