Mixed logical-linear programming (Q1961464): Difference between revisions

Theoretical and algorithmic aspects and applications of mixed logical linear programming are discussed. This is an extension of mixed integer linear programming, and the model has the following form: \[ \min cx\text{ s.t. }p_j(y,h)\to (A^jx\geq a^j),\;j\in J,\;q_i(y,h),\;i\in I. \] \(y\) is a set of logical variables and \(h\) is a set of discrete variables, combined by logical and arithmetical operations respectively. As an example \(q_i(y,h)\) could be \((y_1\vee y_2)\wedge(h_1\neq h_2)\). A system \(A^jx\geq a^j\) is enforced when \(p_j(y,h)\) is true, and this is the continuous part. \((x,y,h)\) is then feasible if \((y,h)\) makes all \(q_i(y,h)\) true and \(x\) satisfies the linear systems corresponding to the formulas \(p_j(y,h)\) that \((y,h)\) makes true.