ParametricMCP represents a mixed complementarity problem (MCP) parameterized by some vector θ.

Solutions of this problem solve f(z, θ) ⟂ lb ≤ z ≤ ub.

The main constructor for compiling a ParametricMCP from

• f: callabale as f(z, θ) that maps a lenght n vector of decision variables z and a parameter vector θ of size parameter_dimension to an lenght n vector output.
• lower_bounds: A lenght n vector of element-wise lower bounds on the decision variables z.
• upper_bounds: A length n vector of element-wise upper bounds on the decision variables z.
• parameter_dimension: the size of the parameter vector θ in f.

Note, this constructor uses Symbolics.jl to compile the relevant low-level functions. Therefore, f must be implemented in a sufficiently generic way that supports symbolic evaluation. In cases where that is strictly infeasible, you can still use the low-level constructor to generate a ParametricMCP. In general, however, the use of this convenience construtor is advised.

Symbolic version of the ParmetricMCP constructor. If you have f and z already in terms of symbolic variables, use this.

Convert a Julia sparse array M into the COO format required by PATH. This implementation has been extracted from here

Get the (sparse) linear indices of all entries that are constant in the symbolic matrix M w.r.t. symbolic vector z.

Returns the number of decision variables for this problem.

Returns the number of decision variables for this problem.