Algorithm Comparison Chart

PHX Optimizers

Algorithm Multi-Objective Max # of Objectives Discrete Variables Num of function evaluations for convergence Constraints Non-smooth responses Minimize or Maximize Solve for Handles failed runs
OPTLIB Gradient Optimizer false 1 false very true false true false false
Darwin true -1 true many true true true false true

DOT

Algorithm Multi-Objective Max # of Objectives Discrete Variables Num of function evaluations for convergence Constraints Non-smooth responses Minimize or Maximize Solve for Handles failed runs
DOT SQP false 1 false small true false true false false
DOT SLP false 1 false small true false true false false
DOT MMFD false 1 false small true false true false false
DOT BFGS false 1 false small false false true false false
DOT FR false 1 false small false false true false false
BIGDOT false 1 false small true false true false false

A Collection of Non-gradient Optimizers

Algorithm Multi-Objective Max # of Objectives Discrete Variables Num of function evaluations for convergence Constraints Non-smooth responses Minimize or Maximize Solve for Handles failed runs
Hooke-Jeeves false 1 false moderate false maybe true false true
Nelder-Mead false 1 false moderate false maybe true false false
NSGA II true -1 true many true true true false true
EVOLVE false 1 true many true true true false false

DAKOTA

Algorithm Multi-Objective Max # of Objectives Discrete Variables Num of function evaluations for convergence Constraints Non-smooth responses Minimize or Maximize Solve for Handles failed runs
DAKOTA APPS false 1 false moderate true maybe true false false
DAKOTA Coliny DIRECT false 1 false moderate true (fixed penalty) true true false false
DAKOTA EA false 1 true many true (penalty function) true true false false
DAKOTA Coliny PS false 1 false moderate true (adaptive penalty function) maybe true false false
DAKOTA Coliny SW false 1 false moderate true (adaptive penalty function) maybe true false false
DAKOTA CONMIN false 1 false very true false true false false
DAKOTA MOGA true -1 true many true true true false false
DAKOTA NCSU DIRECT false 1 false moderate false true true false false
DAKOTA OPT++ Finite Difference Newton false 1 false very true false true false false
DAKOTA OPT++ Quasi Newton false 1 false very true false true false false
DAKOTA OPT++ PDS false 1 false very true false true false false