NSGA_II - performances are not compliant to the original implementation by Deb [sf#4]

esa / pagmo

A C++ / Python platform to perform parallel computations of optimisation tasks (global and local) via the asynchronous generalized island model. State of the art optimization algorithms are included. A common interface is provided to other optimization frameworks/algorithms such as NLOPT, SciPy, SNOPT, IPOPT, GSL

GNU General Public License v3.0

268 stars 87 forks source link

NSGA_II - performances are not compliant to the original implementation by Deb [sf#4] #4

Closed darioizzo closed 10 years ago

darioizzo commented 10 years ago

Originally created by darioizzo on 2012-09-04 19:42:02.

The algorithm performances are not equal to those of the original algorithm. Suspect is that crowding distance and crossover operator have slightly different implementations

darioizzo commented 10 years ago

Originally posted by darioizzo on 2012-09-04 19:42:21.

status: open --> accepted

darioizzo commented 10 years ago

Originally posted by mmarcusx on 2012-10-23 15:24:36.

Description has changed:

Diff:


--- old
+++ new
@@ -1 +1 @@
-The algorithm performances are not equal to those of the original algorithm. Suspect is that crowding distance and crossover operator have slightly different imlementations
+The algorithm performances are not equal to those of the original algorithm. Suspect is that crowding distance and crossover operator have slightly different implementations

status: accepted --> closed

darioizzo commented 10 years ago

Originally posted by mmarcusx on 2012-10-23 15:24:36.

bug fixed (In the crossover/mutation methods was a minus instead of a plus at some point). Convergence speed is now similar to the original implementation by Deb. However, the crowding distance is twice as much as the value by Deb. This should not effect the algorithm however, as it just uses the crowding distance to compute an ordering and never looks at the actual values.