crease_ga.adaptation_params

class crease_ga.adaptation_params(gdmmin=0.005, gdmmax=0.85, pcmin=0.1, pcmax=1, pmmin=0.006, pmmax=0.25, kgdm=1.1, pc=0.6, pm=0.001)

Class for all adaptation parameters needed for a GA run.

Attributes:

gdmmin: float. Default=0.005.

The minimum acceptable value of gdm, a measurement of diversity within a generation (high gdm means low diversity, and vice versa). If gdm of the current generation falls below gdmmin, pc will be multiplied by kgdm and pm will be divided by kgdm to reduce diversity.

gdmmax: float. Default=0.85.

The maximum acceptable value of gdm, a measurement of diversity within a generation (high gdm means low diversity, and vice versa). If gdm of the current generation exceeds gdmmax, pc will be divided by kgdm and pm will be multiplied by kgdm to increase diversity.

pcmin: float. Default=0.1.

Minimum value of pc. pc cannot be further adjusted below pcmin, even if gdm is still too high.

pcmax: float. Default=1.

Maximum value of pc. pc cannot be further adjusted above pcmax, even if gdm is still too low.

pmmin: float. Default=0.006.

Minimum value of pm. pm cannot be further adjusted below pmmin, even if gdm is still too low.

pmmin: float. Default=0.25.

Maximum value of pm. pm cannot be further adjusted above pmmax, even if gdm is still too high.

kgdm: float. Default=1.1.

Should be > 1. The magnitude of adjustment for pc and pm in case gdm falls ouside of [ gdmmin,`gdmmax` ].

pc: float. Default=0.6.

possibility of a crossover action happening on an individual in the next generation. pc is updated after each generation according to gdm.

pm: float. Default=0.001.

possibillity of a mutation action happening on each gene in an individual. pm is updated after each generation according to gdm.

update(gdm)

Update pc and pm according to a gdm value.