1Blurb:: 2Define upper bounds for the linear inequality constraint 3Description:: 4In the inequality case, the 5lower \f$a_l\f$ and upper \f$a_u\f$ bounds provide constraint limits for the 6two-sided formulation: 7\f[a_l \leq Ax \leq a_u\f] 8 9Where \f$A\f$ is the constrain matrix of variable coefficients. 10 11As with nonlinear inequality constraints (see \ref responses-objective_functions), the 12default linear inequality constraint bounds are selected so that 13one-sided inequalities of the form 14\f[Ax \leq 0.0\f] 15result when there are no user bounds specifications (this provides 16backwards compatibility with previous %Dakota versions). 17 18In a user 19bounds specification, any upper bound values greater than \c 20+bigRealBoundSize (1.e+30, as defined in Minimizer) are treated as 21+infinity and any lower bound values less than \c -bigRealBoundSize 22are treated as -infinity. 23 24This feature is commonly used to drop one 25of the bounds in order to specify a 1-sided constraint (just as the 26default lower bounds drop out since \c -DBL_MAX < \c 27-bigRealBoundSize). 28 29Topics:: linear_constraints 30Examples:: 31Examples of specifying linear inequality constraints to Dakota are 32provided on the \ref variables-linear_inequality_constraint_matrix page. 33Theory:: 34Faq:: 35See_Also:: 36