void QpToImplicit::init() {
// Call the base class initializer
ImplicitFunctionInternal::init();
// Free variable in the NLP
MX u = MX::sym("u", input(iin_).sparsity());
// So that we can pass it on to createParent
std::vector<Sparsity> sps;
for (int i=0; i<getNumInputs(); ++i)
if (i!=iin_) sps.push_back(input(i).sparsity());
// u groups all parameters in an MX
std::vector< MX > inputs;
MX p = createParent(sps, inputs);
// Dummy NLP objective
MX nlp_f = 0;
// NLP constraints
std::vector< MX > args_call(getNumInputs());
args_call[iin_] = u;
for (int i=0, i2=0; i<getNumInputs(); ++i)
if (i!=iin_) args_call[i] = inputs[i2++];
MX nlp_g = f_.call(args_call).at(iout_);
// We're going to use two-argument objective and constraints to allow the use of parameters
MXFunction nlp(nlpIn("x", u, "p", p), nlpOut("f", nlp_f, "g", nlp_g));
// Create an NlpSolver instance
solver_ = NlpSolver(getOption(solvername()), nlp);
if (hasSetOption(optionsname())) solver_.setOption(getOption(optionsname()));
solver_.init();
}
void QpToImplicit::init() {
// Call the base class initializer
ImplicitFunctionInternal::init();
// Free variable in the NLP
MX u = MX::sym("u", input(iin_).sparsity());
// So that we can pass it on to createParent
std::vector<MX> inputs;
for (int i=0; i<nIn(); ++i) {
if (i!=iin_) {
stringstream ss;
ss << "p" << i;
inputs.push_back(MX::sym(ss.str(), input(i).sparsity()));
}
}
MX p = veccat(inputs);
// Dummy NLP objective
MX nlp_f = 0;
// NLP constraints
std::vector< MX > args_call(nIn());
args_call[iin_] = u;
for (int i=0, i2=0; i<nIn(); ++i)
if (i!=iin_) args_call[i] = inputs[i2++];
MX nlp_g = f_(args_call).at(iout_);
// We're going to use two-argument objective and constraints to allow the use of parameters
MXFunction nlp("nlp", nlpIn("x", u, "p", p), nlpOut("f", nlp_f, "g", nlp_g));
Dict options;
if (hasSetOption(optionsname())) options = getOption(optionsname());
// Create an NlpSolver instance
solver_ = NlpSolver("nlpsolver", getOption(solvername()), nlp, options);
}
