Benchmark Problems

The following table lists all optimization problems mathlayer has been benchmarked against for its optimization algorithms. Its fields are defined as follows:
  • ID: identifier for the optimization problem
  • Objectives: number of objectives
  • Dimension: dimension of the problem
  • Convex: is problem convex (0: no, 1: yes)
  • Smoothness: how many times are the functions continuously differentiable (-1 is used for discontinuous functions)
  • Observations: number of points
They have been splitted into three categories: Single objective minimization [1], Multiobjective problems [2] and Curve fitting problems [3].

Single objective minimization^

ID Dimension Convex Smoothness
Rosenbrock 2 0
Three-Hump Camel 2 1
Beale 2 1
Matyas 2 1
Ackley 2 0 0
Bohachevsky1 2 1
Bohachevsky2 2 1
Booth 2 1
Adjiman 2 0
Bukin6 2 0 0
Sphere 2 1
Easom 2 1
Freudenstein Roth 2 1
Cross-in-tray 2 0
McCormick 2 0
Schaffer2 2 0
Styblinski-Tang 2 0
Levi13 2 0
Eggholder 2 0
Holder table 2 0
Goldstein-Price 2 0
Bird 2 0
Chichinadze 2 0
Drop-Wave 2 0
Rastrigin 2 0
Michalewicz 2 0
Schwefel 2 0
Gramacy-Lee 1 0
Griewank 2 0
^

Multiobjective problems

ID Objectives Dimension Smoothness
Binh-Korn 2 2 0
Kursawe 2 3 0
constr 2 2 -
Fonseca-Fleming 2 3 -
Osyczka-Kundu 2 6 -
SchafferMO1 2 3 -
SchafferMO2 2 1 -
Chakong-Haimes 2 2 -
Viennet 3 2 -
zdt1 2 30 0
zdt2 2 30 -
zdt4 2 10 0
Poloni 2 2 0
Hajime-Yabumoto-Mori-Yoshikazu 2 2 -
ctp1 2 2 -
^

Curve fitting problems

ID Objectives Dimension Observations
Misra1a 1 2 14
Chwirut2 1 3 54
Chwirut1 1 3 214
Lanczos3 1 6 24
Gauss1 1 8 250
Gauss2 1 8 250
DanWood 1 2 6
Misra1b 1 2 14
Kirby2 1 5 151
Hahn1 1 7 236
Nelson 1 3 128
MGH17 1 5 33
Lanczos1 1 6 24
Lanczos2 1 6 24
Gauss3 1 8 250
Misra1c 1 2 14
Misra1d 1 2 14
Roszman1 1 4 25
ENSO 1 9 168
MGH09 1 4 11
Thurber 1 37 7
BoxBOD 1 2 6
Rat42 1 3 9
MGH10 1 3 16
Eckerle4 1 3 35
Rat43 1 4 15
Bennett5 1 3 154
Ratkowsky2 1 3 9