makeED2Function {smoof} R Documentation
ED2 Function
Description
Builds and returns the multi-objective ED2 test problem.
The ED2 test problem is defined as follows:
Minimize f j ( x ) = 1 F n a t m i n ( x ) + 1 ⋅ p ~ ( Θ ( X ) ) f_j(\mathbf{x}) = \frac{1}{F_{natmin}(\mathbf{x}) + 1} \cdot \tilde{p}(\Theta (\mathbf{X})) f j ( x ) = F na t min ( x ) + 1 1 ⋅ p ~ ( Θ ( X ))
, for j = 1 , … , m j = 1, \ldots, m j = 1 , … , m
,
with x = ( x 1 , … , x n ) T \mathbf{x} = (x_1, \ldots, x_n)^T x = ( x 1 , … , x n ) T
, where 0 ≤ x i ≤ 1 0 \leq x_i \leq 1 0 ≤ x i ≤ 1
,
and Θ = ( θ 1 , … , θ m − 1 ) \Theta = (\theta_1, \ldots, \theta_{m-1}) Θ = ( θ 1 , … , θ m − 1 )
,
where 0 ≤ θ j ≤ π 2 0 \le \theta_j \le \frac{\pi}{2} 0 ≤ θ j ≤ 2 π
, for i = 1 , … , n , i = 1, \ldots, n, i = 1 , … , n ,
and j = 1 , … , m − 1 j = 1, \ldots, m - 1 j = 1 , … , m − 1
.
Moreover F n a t m i n ( x ) = b + ( r ( x ) − a ) + 0.5 + 0.5 ⋅ ( 2 π ⋅ ( r ( x ) − a ) + π ) F_{natmin}(\mathbf{x}) = b + (r(\mathbf{x}) - a) + 0.5 + 0.5 \cdot (2 \pi \cdot (r(\mathbf{x}) - a) + \pi) F na t min ( x ) = b + ( r ( x ) − a ) + 0.5 + 0.5 ⋅ ( 2 π ⋅ ( r ( x ) − a ) + π )
with a ≈ 0.051373 a \approx 0.051373 a ≈ 0.051373
, b ≈ 0.0253235 b \approx 0.0253235 b ≈ 0.0253235
, and r ( X ) = x m 2 + … , x n 2 r(\mathbf{X}) = \sqrt{x_m^2 + \ldots, x_n^2} r ( X ) = x m 2 + … , x n 2
, as well as
p ~ 1 ( Θ ) = cos ( θ 1 ) 2 / γ \tilde{p}_1(\Theta) = \cos(\theta_1)^{2/\gamma} p ~ 1 ( Θ ) = cos ( θ 1 ) 2/ γ
,
p ~ j ( Θ ) = ( sin ( θ 1 ) ⋅ … ⋅ sin ( θ j − 1 ) ⋅ cos ( θ j ) ) 2 / γ \tilde{p}_j(\Theta) = \left( \sin(\theta_1) \cdot \ldots \cdot \sin(\theta_{j - 1}) \cdot \cos(\theta_j) \right)^{2/\gamma} p ~ j ( Θ ) = ( sin ( θ 1 ) ⋅ … ⋅ sin ( θ j − 1 ) ⋅ cos ( θ j ) ) 2/ γ
,
for 2 ≤ j ≤ m − 1 2 \le j \le m - 1 2 ≤ j ≤ m − 1
,
and p ~ m ( Θ ) = ( sin ( θ 1 ) ⋅ … ⋅ sin ( θ m − 1 ) ) 2 / γ \tilde{p}_m(\Theta) = \left( \sin(\theta_1) \cdot \ldots \cdot \sin(\theta_{m - 1}) \right)^{2/\gamma} p ~ m ( Θ ) = ( sin ( θ 1 ) ⋅ … ⋅ sin ( θ m − 1 ) ) 2/ γ
.
Usage
makeED2Function(dimensions, n.objectives, gamma = 2, theta)
Arguments
dimensions
[integer(1)
]
Number of decision variables.
n.objectives
[integer(1)
]
Number of objectives.
gamma
[numeric(1)
]
Optional parameter. Default is 2, which is recommended by Emmerich and Deutz.
theta
[numeric(dimensions)
]
Parameter vector, whose components have to be between 0
and 0.5*pi
.
The default is theta = (pi/2) * x
(with x
being the point from the decision space) as recommended by Emmerich and Deutz.
Value
[smoof_multi_objective_function
]
References
M. T. M. Emmerich and A. H. Deutz. Test Problems based on Lame
Superspheres. Proceedings of the International Conference on Evolutionary
Multi-Criterion Optimization (EMO 2007), pp. 922-936, Springer, 2007.
[Package
smoof version 1.6.0.3
Index ]