1# DAKOTA INPUT FILE - dakota_sbo_illum.in 2 3# Demonstrates the use of approximation models and a trust region 4# optimization environment in the illumination example problem. 5 6environment, 7 graphics 8 method_pointer = 'SBLO' 9 10method, 11 id_method = 'SBLO' 12 surrogate_based_local 13 model_pointer = 'SURROGATE' 14 approx_method_pointer = 'NLP' 15 max_iterations = 100, 16 trust_region 17 initial_size = 0.10 18 contraction_factor = 0.50 19 expansion_factor = 1.50 20 21method, 22 id_method = 'NLP' 23# optpp_newton, 24# optpp_cg, 25# npsol, 26# dot_bfgs, 27# dot_frcg, 28 conmin_frcg 29 max_iterations = 50, 30 convergence_tolerance = 1e-8 31 32model, 33 id_model = 'SURROGATE' 34 surrogate global 35 responses_pointer = 'SURROGATE_RESP' 36 dace_method_pointer = 'SAMPLING' 37# reuse_samples region 38# use_derivatives #s2 39 correction multiplicative zeroth_order 40# neural_network 41# polynomial quadratic 42 gaussian_process surfpack 43 correlation_lengths = 0.707106781186547 0.707106781186547 0.707106781186547 0.707106781186547 0.707106781186547 0.707106781186547 0.707106781186547 #s0 44 trend constant 45# mars 46# surogate local taylor_series 47# actual_model_pointer = 'TRUTH' 48 49variables, 50 continuous_design = 7 51 initial_point .5 .5 .5 .5 .5 .5 .5 52 lower_bounds 0. 0. 0. 0. 0. 0. 0. 53 upper_bounds 1. 1. 1. 1. 1. 1. 1. 54 descriptors 'x1' 'x2' 'x3' 'x4' 'x5' 'x6' 'x7' 55 56responses, 57 id_responses = 'SURROGATE_RESP' 58 objective_functions = 1 59 analytic_gradients 60# numerical_gradients 61# method_source dakota 62# interval_type central 63# fd_gradient_step_size = .00001 64# analytic_hessians 65 no_hessians 66 67############################################### 68# SAMPLING method specifications for building # 69# surrogate functions # 70############################################### 71method, 72 id_method = 'SAMPLING' 73 model_pointer = 'TRUTH' 74# dace box_behnken 75# dace central_composite 76 dace lhs 77 seed = 13579 78 samples = 40 79# dace oas seed = 5 80# samples = 49 symbols = 7 81 82model, 83 id_model = 'TRUTH' 84 single 85 interface_pointer = 'TRUE_FN' 86 responses_pointer = 'TRUE_RESP' 87 88interface, 89 direct 90 id_interface = 'TRUE_FN' 91 analysis_drivers = 'illumination' 92 93responses, 94 id_responses = 'TRUE_RESP' 95 objective_functions = 1 96# analytic_gradients #s2 97 no_gradients #s0,#s1 98# numerical_gradients 99# method_source dakota 100# interval_type central 101# fd_gradient_step_size = .0001 102 no_hessians 103