Mathematical Modelling of a Friction Stir Welding Process to Predict the Joint Strength of Two Dissimilar Aluminium Alloys Using Experimental Data and Genetic Programming
Table 3
Parameter setting for genetic programming.
Parameters
Value assigned
Population size ()
600
Number of generations
1000
Maximum depth of tree
6
Maximum generation
60
Functional set
Multiply (×), plus (+), minus (−), divide (/)
Terminal set
Tilt angle, axial load, tool rotational speed, welding speed, thickness of plate, tensile strength, elongation, and impact strength.
Number of runs
100
Mutation rate
15%
Crossover rate
(i) 0.80 for nonhomologous (ii) 0.20 for homologous
Reproduction rate
0.05
Fitness
Termination
Individual program emerges when the sum of absolute errors is zero (is less than specified) (a) the required number of runs is accomplished (10000) or (b) the required correlation coefficient () is obtained (when is less than 0.1)