OMCompiler/SimulationRuntime/c/optimization/eval_all/EvalF.c
| Line | Branch | Exec | Source |
|---|---|---|---|
| 1 | /* | ||
| 2 | * This file belongs to the OpenModelica Run-Time System | ||
| 3 | * | ||
| 4 | * Copyright (c) 1998-2026, Open Source Modelica Consortium (OSMC), c/o Linköpings | ||
| 5 | * universitet, Department of Computer and Information Science, SE-58183 Linköping, Sweden. All rights | ||
| 6 | * reserved. | ||
| 7 | * | ||
| 8 | * THIS PROGRAM IS PROVIDED UNDER THE TERMS OF THE BSD NEW LICENSE OR THE | ||
| 9 | * AGPL VERSION 3 LICENSE OR THE OSMC PUBLIC LICENSE (OSMC-PL) VERSION 1.8. ANY | ||
| 10 | * USE, REPRODUCTION OR DISTRIBUTION OF THIS PROGRAM CONSTITUTES RECIPIENT'S | ||
| 11 | * ACCEPTANCE OF THE BSD NEW LICENSE OR THE OSMC PUBLIC LICENSE OR THE AGPL | ||
| 12 | * VERSION 3, ACCORDING TO RECIPIENTS CHOICE. | ||
| 13 | * | ||
| 14 | * The OpenModelica software and the OSMC (Open Source Modelica Consortium) Public License | ||
| 15 | * (OSMC-PL) are obtained from OSMC, either from the above address, from the URLs: | ||
| 16 | * http://www.openmodelica.org or https://github.com/OpenModelica/ or | ||
| 17 | * http://www.ida.liu.se/projects/OpenModelica, and in the OpenModelica distribution. GNU | ||
| 18 | * AGPL version 3 is obtained from: https://www.gnu.org/licenses/licenses.html#GPL. The BSD NEW | ||
| 19 | * License is obtained from: http://www.opensource.org/licenses/BSD-3-Clause. | ||
| 20 | * | ||
| 21 | * This program is distributed WITHOUT ANY WARRANTY; without even the implied warranty of | ||
| 22 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE, EXCEPT AS EXPRESSLY | ||
| 23 | * SET FORTH IN THE BY RECIPIENT SELECTED SUBSIDIARY LICENSE CONDITIONS OF | ||
| 24 | * OSMC-PL. | ||
| 25 | * | ||
| 26 | */ | ||
| 27 | |||
| 28 | /*! EvalF.c | ||
| 29 | */ | ||
| 30 | |||
| 31 | #include "../OptimizerData.h" | ||
| 32 | #include "../OptimizerLocalFunction.h" | ||
| 33 | |||
| 34 | |||
| 35 | /* eval object function | ||
| 36 | */ | ||
| 37 | ✗ | Bool evalfF(ipindex n, ipnumber * vopt, Bool new_x, ipnumber *objValue, void * useData){ | |
| 38 | |||
| 39 | OptData *optData = (OptData*)useData; | ||
| 40 | |||
| 41 | ✗ | const modelica_boolean la = optData->s.lagrange; | |
| 42 | ✗ | const modelica_boolean ma = optData->s.mayer; | |
| 43 | |||
| 44 | long double mayer = 0.0; | ||
| 45 | long double lagrange = 0.0; | ||
| 46 | |||
| 47 | ✗ | if(new_x) | |
| 48 | ✗ | optData2ModelData(optData, vopt, 1); | |
| 49 | |||
| 50 | ✗ | if(la){ | |
| 51 | ✗ | const int nsi = optData->dim.nsi; | |
| 52 | ✗ | const int np = optData->dim.np; | |
| 53 | ✗ | const int il = optData->dim.index_lagrange; | |
| 54 | |||
| 55 | ✗ | const long double * const b = optData->rk.b; | |
| 56 | ✗ | const long double * const dt = optData->time.dt; | |
| 57 | |||
| 58 | ✗ | modelica_real *** v = optData->v; | |
| 59 | ✗ | long double *erg = (long double*)malloc(np * sizeof(long double)); | |
| 60 | int i,j; | ||
| 61 | |||
| 62 | ✗ | for(j = 0; j< np; ++j){ | |
| 63 | ✗ | erg[j] = dt[0]*v[0][j][il]; | |
| 64 | } | ||
| 65 | |||
| 66 | ✗ | for(i = 1; i < nsi; ++i){ | |
| 67 | ✗ | for(j = 0; j< np; ++j){ | |
| 68 | ✗ | erg[j] += dt[i]*v[i][j][il]; | |
| 69 | } | ||
| 70 | } | ||
| 71 | |||
| 72 | ✗ | for(j = 0; j< np; ++j) | |
| 73 | ✗ | lagrange += b[j]*erg[j]; | |
| 74 | ✗ | free(erg); | |
| 75 | } | ||
| 76 | |||
| 77 | ✗ | if(ma){ | |
| 78 | ✗ | modelica_real *** v = optData->v; | |
| 79 | ✗ | const int nsi = optData->dim.nsi; | |
| 80 | ✗ | const int np = optData->dim.np; | |
| 81 | ✗ | const int im = optData->dim.index_mayer; | |
| 82 | ✗ | mayer = v[nsi-1][np-1][im]; | |
| 83 | } | ||
| 84 | |||
| 85 | ✗ | *objValue = (ipnumber)(lagrange + mayer); | |
| 86 | |||
| 87 | ✗ | return TRUE; | |
| 88 | } | ||
| 89 | |||
| 90 | |||
| 91 | /*! | ||
| 92 | * eval derivation (object func) | ||
| 93 | * author: Vitalij Ruge | ||
| 94 | **/ | ||
| 95 | ✗ | Bool evalfDiffF(ipindex n, double * vopt, Bool new_x, ipnumber *gradF, void * useData){ | |
| 96 | OptData *optData = (OptData*)useData; | ||
| 97 | |||
| 98 | ✗ | const int nv = optData->dim.nv; | |
| 99 | ✗ | const int nsi = optData->dim.nsi; | |
| 100 | ✗ | const int np = optData->dim.np; | |
| 101 | ✗ | const int nJ = optData->dim.nJ; | |
| 102 | ✗ | const int nJ1 = optData->dim.nJ + 1; | |
| 103 | |||
| 104 | ✗ | const modelica_boolean la = optData->s.lagrange; | |
| 105 | ✗ | const modelica_boolean ma = optData->s.mayer; | |
| 106 | |||
| 107 | ✗ | if(new_x) | |
| 108 | ✗ | optData2ModelData(optData, vopt, 1); | |
| 109 | |||
| 110 | ✗ | if(la){ | |
| 111 | |||
| 112 | int i, j, ii; | ||
| 113 | modelica_real * gradL; | ||
| 114 | |||
| 115 | ✗ | for(i = 0, ii = 0; i < nsi - 1; ++i){ | |
| 116 | ✗ | for(j = 0; j < np; ++j, ii += nv){ | |
| 117 | ✗ | gradL = optData->J[i][j][nJ]; | |
| 118 | ✗ | memcpy(gradF + ii, gradL, nv*sizeof(modelica_real)); | |
| 119 | } | ||
| 120 | } | ||
| 121 | |||
| 122 | ✗ | for(j = 0; j < np; ++j, ii += nv){ | |
| 123 | ✗ | gradL = optData->J[i][j][nJ];; | |
| 124 | ✗ | memcpy(gradF + ii, gradL, nv*sizeof(modelica_real)); | |
| 125 | } | ||
| 126 | |||
| 127 | }else{ | ||
| 128 | ✗ | memset(gradF,0.0,n*sizeof(ipnumber)); | |
| 129 | } | ||
| 130 | |||
| 131 | ✗ | if(ma){ | |
| 132 | ✗ | modelica_real * gradM = optData->J[nsi - 1][np -1][nJ1]; | |
| 133 | ✗ | if(la){ | |
| 134 | int i; | ||
| 135 | ✗ | const int nnv = n - nv; | |
| 136 | ✗ | for(i = 0; i < nv; ++i) | |
| 137 | ✗ | gradF[nnv + i] += gradM[i]; | |
| 138 | }else{ | ||
| 139 | ✗ | memcpy(gradF + n - nv, gradM, nv*sizeof(modelica_real)); | |
| 140 | } | ||
| 141 | |||
| 142 | } | ||
| 143 | |||
| 144 | ✗ | return TRUE; | |
| 145 | } | ||
| 146 |