Macro grb::c

c!() { /* proc-macro */ }

A proc-macro for creating constraint objects.

Syntax

Inequality constraints

To create an IneqExpr object for a linear or quadratic constraint, the syntax is

c!( LHS CMP RHS )

LHS and RHS should be valid algebraic expressions involving Var objects and numeric constants. For example, if x, y and z are Var objects and vars is an Vec<Var> objects, these are valid:

  c!(vars.iter().grb_sum() == x );
  c!( x + 1/2 == 1.4*y - 2*z );
  c!( 2*x >= z*y );
  c!( 2*x >= 7*(z*y) ); // note the brackets on the non-linear term when a coefficient is present

but the following are not:

  c!(vars.iter().sum() == x ); // cannot infer type on sum() call
  c!( 2*x >= z >= y ); // chained comparison
  c!( 2*x >= 7*z*y ); // no brackets around var*var when a coefficient is present

The macro expands c!( LHS == RHS ) to:

grb::constr::IneqExpr {
  lhs: grb::Expr::from(LHS),
  sense: grb::ConstrSense::Equal,
  rhs: grb::Expr::from(RHS),
};

Range constraints

To create a RangeExpr object for a range constraint, use the syntax

c!( EXPR in LB..UB )
c!( EXPR in LB.. )
c!( EXPR in ..UB )
c!( EXPR in .. )

where EXPR is a valid expression, like LHS and RHS above. Additionally, EXPR must be linear, although this is not checked at compile-time.

LB and UB can be any expression that evaluates to type that can be cast to a f64 using the as operator. For example, the following are valid (variables have the meaning as above):

  c!( x - y + 2*z in 0..200 );
  c!( x - y + 2*z in 1.. );
  c!( x - y in (1.0/3.0)..(1<<4));