Extending SCIP for solving mixed-integer nonlinear programs

Size: px
Start display at page:

Download "Extending SCIP for solving mixed-integer nonlinear programs"

Transcription

1 Extending SCIP for solving mixed-integer nonlinear programs Stefan Vigerske Humboldt-Universität zu Berlin DFG Research Center MATHEON Mathematics for key technologies Spring Workshop on Computational Issues in Mixed Integer Nonlinear Programming

2 MINLP project in Berlin: cooperation of Humboldt University (Römisch, V.) and Zuse Institute (Grötschel, Bley, Gleixner, Koch, Pfetsch) started October 2008 work in progress joint work with T. Gellermann (TU Darmstadt), A. Gleixner (ZIB), T. Koch (ZIB), A. Neumaier (Vienna), M. Pfetsch (TU Braunschweig) Introduction Extending SCIP for solving mixed-integer nonlinear programs 2 / 31

3 MINLP project in Berlin: cooperation of Humboldt University (Römisch, V.) and Zuse Institute (Grötschel, Bley, Gleixner, Koch, Pfetsch) started October 2008 work in progress joint work with T. Gellermann (TU Darmstadt), A. Gleixner (ZIB), T. Koch (ZIB), A. Neumaier (Vienna), M. Pfetsch (TU Braunschweig) Introduction Aim: teach Constraint-Integer-Programming solver SCIP how to handle nonlinear constraints LP-based branch-and-cut Extending SCIP for solving mixed-integer nonlinear programs 2 / 31

4 MINLP project in Berlin: cooperation of Humboldt University (Römisch, V.) and Zuse Institute (Grötschel, Bley, Gleixner, Koch, Pfetsch) started October 2008 work in progress joint work with T. Gellermann (TU Darmstadt), A. Gleixner (ZIB), T. Koch (ZIB), A. Neumaier (Vienna), M. Pfetsch (TU Braunschweig) Introduction Aim: teach Constraint-Integer-Programming solver SCIP how to handle nonlinear constraints LP-based branch-and-cut start with quadratic problems extend step-by-step to other types of constraints Extending SCIP for solving mixed-integer nonlinear programs 2 / 31

5 hybrid estim estimate dfs spx qso restart dfs bfs xprs none implics dualfix bound shift msk intto binary Node selector default default LP probing Presolver cpx clp trivial Event Display default rounding rins pscost diving rootsol diving rens Impli cations Dialog simple rounding shifting oneopt fix cnf octane cip lp Tree Cutpool veclen diving ccg mps objpscost diving Reader Heuristic mutation opb zpl Conflict actcons diving local branching ppm sos Pricer rlp coef diving sol int shifting linesearch diving cross over dins cmir clique fixand infer guided diving flow cover Propa gator intdiving Variable feaspump gomory pseudo obj Relaxer fracdiving Separator implied bounds zero half root redcost setppc or relps cost sos1 logicor intobj strong cg random mcf redcost sos2 linear Branch var bound and Constraint Handler knap sack allfull strong pscost full strong mostinf bound xor disjunc. integral in ference leastinf count sols indi cator SCIP is a Branch-Cut-Price framework for Solving Constraint Integer Programs SCIP developed by T. Achterberg, T. Berthold, G. Gamrath, S. Heinz, T. Koch, A. Martin, M. Pfetsch, C. Raack, R. Waniek, M. Winkler, K. Wolter SCIP Extending SCIP for solving mixed-integer nonlinear programs 3 / 31

6 hybrid estim estimate dfs spx qso restart dfs bfs xprs none implics dualfix bound shift msk intto binary Node selector default default LP probing Presolver cpx clp trivial Event Display default rounding rins pscost diving rootsol diving rens Impli cations Dialog simple rounding shifting oneopt fix cnf octane cip lp Tree Cutpool veclen diving ccg mps objpscost diving Reader Heuristic mutation opb zpl Conflict actcons diving local branching ppm sos Pricer rlp coef diving sol int shifting linesearch diving cross over dins cmir clique fixand infer guided diving flow cover Propa gator intdiving Variable feaspump gomory pseudo obj Relaxer fracdiving Separator implied bounds zero half root redcost setppc or relps cost sos1 logicor intobj strong cg random mcf redcost sos2 linear Branch var bound and Constraint Handler knap sack allfull strong pscost full strong mostinf bound xor disjunc. integral in ference leastinf count sols indi cator SCIP is a Branch-Cut-Price framework for Solving Constraint Integer Programs SCIP developed by T. Achterberg, T. Berthold, G. Gamrath, S. Heinz, T. Koch, A. Martin, M. Pfetsch, C. Raack, R. Waniek, M. Winkler, K. Wolter constraint oriented everything is a plugin SCIP Extending SCIP for solving mixed-integer nonlinear programs 3 / 31

7 hybrid estim estimate dfs spx qso restart dfs bfs xprs none implics dualfix bound shift msk intto binary Node selector default default LP probing Presolver cpx clp trivial Event Display default rounding rins pscost diving rootsol diving rens Impli cations Dialog simple rounding shifting oneopt fix cnf octane cip lp Tree Cutpool veclen diving ccg mps objpscost diving Reader Heuristic mutation opb zpl Conflict actcons diving local branching ppm sos Pricer rlp coef diving sol int shifting linesearch diving cross over dins cmir clique fixand infer guided diving flow cover Propa gator intdiving Variable feaspump gomory pseudo obj Relaxer fracdiving Separator implied bounds zero half root redcost setppc or relps cost sos1 logicor intobj strong cg random mcf redcost sos2 linear Branch var bound and Constraint Handler knap sack allfull strong pscost full strong mostinf bound xor disjunc. integral in ference leastinf count sols indi cator SCIP is a Branch-Cut-Price framework for Solving Constraint Integer Programs SCIP developed by T. Achterberg, T. Berthold, G. Gamrath, S. Heinz, T. Koch, A. Martin, M. Pfetsch, C. Raack, R. Waniek, M. Winkler, K. Wolter constraint oriented everything is a plugin can be used as very efficient MIP solver strong preprocessing, branching rules, domain propagators, many heuristics and separators SCIP Extending SCIP for solving mixed-integer nonlinear programs 3 / 31

8 hybrid estim estimate dfs spx qso restart dfs bfs xprs none implics dualfix bound shift msk intto binary Node selector default default LP probing Presolver cpx clp trivial Event Display default rounding rins pscost diving rootsol diving rens Impli cations Dialog simple rounding shifting oneopt fix cnf octane cip lp Tree Cutpool veclen diving ccg mps objpscost diving Reader Heuristic mutation opb zpl Conflict actcons diving local branching ppm sos Pricer rlp coef diving sol int shifting linesearch diving cross over dins cmir clique fixand infer guided diving flow cover Propa gator intdiving Variable feaspump gomory pseudo obj Relaxer fracdiving Separator implied bounds zero half root redcost setppc or relps cost sos1 logicor intobj strong cg random mcf redcost sos2 linear Branch var bound and Constraint Handler knap sack allfull strong pscost full strong mostinf bound xor disjunc. integral in ference leastinf count sols indi cator SCIP is a Branch-Cut-Price framework for Solving Constraint Integer Programs SCIP developed by T. Achterberg, T. Berthold, G. Gamrath, S. Heinz, T. Koch, A. Martin, M. Pfetsch, C. Raack, R. Waniek, M. Winkler, K. Wolter constraint oriented everything is a plugin can be used as very efficient MIP solver strong preprocessing, branching rules, domain propagators, many heuristics and separators free for academic use SCIP Extending SCIP for solving mixed-integer nonlinear programs 3 / 31

9 Overview 1 Introduction 2 Quadratic Constraints 3 Nonlinear Constraints (MINLP) 4 Heuristic 5 Preliminary Numerical Results 6 Latest Work Extending SCIP for solving mixed-integer nonlinear programs 4 / 31

10 Overview 1 Introduction 2 Quadratic Constraints 3 Nonlinear Constraints (MINLP) 4 Heuristic 5 Preliminary Numerical Results 6 Latest Work Extending SCIP for solving mixed-integer nonlinear programs 5 / 31

11 Quadratic Constraints: Presolve Presolve: l x Ax + b x u upgrade to linear constraint if possible, substitute {0, 1} 2 by {0, 1},... Extending SCIP for solving mixed-integer nonlinear programs 6 / 31

12 Quadratic Constraints: Presolve l x Ax + b x u Presolve: upgrade to linear constraint if possible, substitute {0, 1} 2 by {0, 1},... find block structure, i.e., partition (J k ) k of {1,..., n} s.t.: l k x J k A k x Jk + b x u and disaggregate l k z k + b x u, z k = x J k A k x Jk Extending SCIP for solving mixed-integer nonlinear programs 6 / 31

13 Quadratic Constraints: Presolve l x Ax + b x u Presolve: upgrade to linear constraint if possible, substitute {0, 1} 2 by {0, 1},... find block structure, i.e., partition (J k ) k of {1,..., n} s.t.: l k x J k A k x Jk + b x u and disaggregate l k z k + b x u, z k = x J k A k x Jk domain reduction Extending SCIP for solving mixed-integer nonlinear programs 6 / 31

14 Quadratic Constraints: Presolve l x Ax + b x u Presolve: upgrade to linear constraint if possible, substitute {0, 1} 2 by {0, 1},... find block structure, i.e., partition (J k ) k of {1,..., n} s.t.: l k x J k A k x Jk + b x u and disaggregate l k z k + b x u, z k = x J k A k x Jk domain reduction compute min/max eigenvalues for each A k Extending SCIP for solving mixed-integer nonlinear programs 6 / 31

15 Quadratic Constraints: Separation xj k A k x Jk + b x u Separation: try to cut off LP solution by a linear cut k Extending SCIP for solving mixed-integer nonlinear programs 7 / 31

16 Quadratic Constraints: Separation xj k A k x Jk + b x u Separation: try to cut off LP solution by a linear cut linearize if convex (min. eigenvalue of A k 0) k Extending SCIP for solving mixed-integer nonlinear programs 7 / 31

17 Quadratic Constraints: Separation xj k A k x Jk + b x u Separation: try to cut off LP solution by a linear cut linearize if convex (min. eigenvalue of A k 0) k apply McCormick for each bilinear term, x i x j x L i x j + x L j x i x L i x L j x i x j x U i x j + x U j x i x U i x U j secant underestimator for concave terms ( x 2 ) Extending SCIP for solving mixed-integer nonlinear programs 7 / 31

18 Quadratic Constraints: Propagation Univariate case: [Domes and Neumaier 2008] ax 2 + bx [l, u] Extending SCIP for solving mixed-integer nonlinear programs 8 / 31

19 Quadratic Constraints: Propagation Univariate case: [Domes and Neumaier 2008] ax 2 + bx [l, u] forward propagation: compute [ l, ū] := {a x 2 + b x : x [x L, x U ]} [l, u] backward propagation: compute {x : a x 2 + b x [ l, ū]} Extending SCIP for solving mixed-integer nonlinear programs 8 / 31

20 Quadratic Constraints: Propagation Separable case: [Domes and Neumaier 2008] a k xk 2 + b kx k [l, u] k Extending SCIP for solving mixed-integer nonlinear programs 8 / 31

21 Quadratic Constraints: Propagation Separable case: [Domes and Neumaier 2008] a k xk 2 + b kx k [l, u] k forward propagation: compute [l k, u k ] := {a k xk 2 + b k x k : x k [xk L, x k U ]} backward propagation: compute {x k : a k xk 2 + b k x k [l, u] j, u j ]} j k[l Extending SCIP for solving mixed-integer nonlinear programs 8 / 31

22 Quadratic Constraints: Propagation Non-separable case: [Domes and Neumaier 2008] a k,k xk 2 + (b k + a k,j x j )x k [l, u] k j:j k Extending SCIP for solving mixed-integer nonlinear programs 8 / 31

23 Quadratic Constraints: Propagation Non-separable case: [Domes and Neumaier 2008] a k,k xk 2 + (b k + a k,j x j )x k [l, u] k j:j k forward propagation: compute [l k, u k ] := {a k,k xk 2 +(b k+ a k,j [xj L, xj U ]) x k : x k [xk L, x k U ]} j:j k backward propagation: compute {x k : a k,k xk 2 +(b k+ a k,j [xj L, xj U ]) x k [l, u] [l k, u k ]} j:j k j:j k Extending SCIP for solving mixed-integer nonlinear programs 8 / 31

24 Overview 1 Introduction 2 Quadratic Constraints 3 Nonlinear Constraints (MINLP) 4 Heuristic 5 Preliminary Numerical Results 6 Latest Work Extending SCIP for solving mixed-integer nonlinear programs 9 / 31

25 Nonlinear Constraints: Presolve Presolve: Given constraint h(x) [l, u], h C 2 (R n, R) Extending SCIP for solving mixed-integer nonlinear programs 10 / 31

26 Nonlinear Constraints: Presolve Presolve: Given constraint h(x) [l, u], h C 2 (R n, R) upgrade to linear or quadratic constraint, if possible Extending SCIP for solving mixed-integer nonlinear programs 10 / 31

27 Nonlinear Constraints: Presolve Presolve: Given constraint h(x) [l, u], h C 2 (R n, R) upgrade to linear or quadratic constraint, if possible find block structure and quadratic terms: h(x) = b x + k x Q k A k x Qk + r g r (x Nr ) with disjoint subsets Q k and N r of {1,..., n} disaggregate as h(x) = b x + k z k + r z r, z k = x Q k A k x Qk, z r = g r (x Nr ) Extending SCIP for solving mixed-integer nonlinear programs 10 / 31

28 Nonlinear Constraints: Presolve Presolve: Given constraint h(x) [l, u], h C 2 (R n, R) upgrade to linear or quadratic constraint, if possible find block structure and quadratic terms: h(x) = b x + k x Q k A k x Qk + r g r (x Nr ) with disjoint subsets Q k and N r of {1,..., n} disaggregate as h(x) = b x + k z k + r z r, z k = x Q k A k x Qk, z r = g r (x Nr ) Check probable convexity/concavity of each g r : compute sign of eigenvalues of Hessian 2 g r in sample points Extending SCIP for solving mixed-integer nonlinear programs 10 / 31

29 Nonlinear Constraints: Separation Separation (overview): Given constraint g(x) 0, g C 2 (R n, R). Solution ˆx of LP relaxation such that g(ˆx) > 0. Extending SCIP for solving mixed-integer nonlinear programs 11 / 31

30 Nonlinear Constraints: Separation Separation (overview): Given constraint g(x) 0, g C 2 (R n, R). Solution ˆx of LP relaxation such that g(ˆx) > 0. If g( ) is convex, linearize at ˆx: g(ˆx) + g(ˆx)(x ˆx) 0 Extending SCIP for solving mixed-integer nonlinear programs 11 / 31

31 Nonlinear Constraints: Separation Separation (overview): Given constraint g(x) 0, g C 2 (R n, R). Solution ˆx of LP relaxation such that g(ˆx) > 0. If g( ) is convex, linearize at ˆx: g(ˆx) + g(ˆx)(x ˆx) 0 If g( ) is not convex, generate interval gradient cut: g( x) + d L i (x i x i ) + d U i (x i x i ) 0 i: x i =x L i i: x i =x U i Extending SCIP for solving mixed-integer nonlinear programs 11 / 31

32 Nonlinear Constraints: Separation Separation (overview): Given constraint g(x) 0, g C 2 (R n, R). Solution ˆx of LP relaxation such that g(ˆx) > 0. If g( ) is convex, linearize at ˆx: g(ˆx) + g(ˆx)(x ˆx) 0 If g( ) is not convex, generate interval gradient cut: g( x) + d L i (x i x i ) + d U i (x i x i ) 0 i: x i =x L i i: x i =x U i If ˆx is not cut off, generate quadratic cut and add as new quadratic constraint: x Ax + b x + c 0 Extending SCIP for solving mixed-integer nonlinear programs 11 / 31

33 Nonlinear Constraints: Interval Gradient Cuts Given function g(x) and box [x L, x U ]. Compute interval gradient (using CppAD) [d L, d U ] = g([x L, x U ]), i.e., g(x) [d L, d U ] for all x [x L, x U ]. Extending SCIP for solving mixed-integer nonlinear programs 12 / 31

34 Nonlinear Constraints: Interval Gradient Cuts Given function g(x) and box [x L, x U ]. Compute interval gradient (using CppAD) [d L, d U ] = g([x L, x U ]), i.e., g(x) [d L, d U ] for all x [x L, x U ]. Let ˆx [x L, x U ]. Then g(ˆx) + min d [d L,d U ] d (x ˆx) g(x) Extending SCIP for solving mixed-integer nonlinear programs 12 / 31

35 Nonlinear Constraints: Interval Gradient Cuts Given function g(x) and box [x L, x U ]. Compute interval gradient (using CppAD) [d L, d U ] = g([x L, x U ]), i.e., g(x) [d L, d U ] for all x [x L, x U ]. Let ˆx [x L, x U ]. Then g(ˆx) + i:x i ˆx i d L i (x i ˆx i ) + di U i:x i <ˆx i (x i ˆx i ) g(x) Extending SCIP for solving mixed-integer nonlinear programs 12 / 31

36 Nonlinear Constraints: Interval Gradient Cuts Given function g(x) and box [x L, x U ]. Compute interval gradient (using CppAD) [d L, d U ] = g([x L, x U ]), i.e., g(x) [d L, d U ] for all x [x L, x U ]. Let ˆx [x L, x U ]. Then g(ˆx) + i:x i ˆx i d L i (x i ˆx i ) + di U i:x i <ˆx i Simple version (implemented): move ˆx to closest vertex of box (ˆx i {x L i, x U i }) g(ˆx) + d L i (x i ˆx i ) + di U (x i ˆx i ) g(x) (x i ˆx i ) g(x) i:ˆx i =x L i i:ˆx i =x U i Extending SCIP for solving mixed-integer nonlinear programs 12 / 31

37 Nonlinear Constraints: Interval Gradient Cuts Given function g(x) and box [x L, x U ]. Compute interval gradient (using CppAD) [d L, d U ] = g([x L, x U ]), i.e., g(x) [d L, d U ] for all x [x L, x U ]. Let ˆx [x L, x U ]. Then g(ˆx) + i:x i ˆx i d L i (x i ˆx i ) + di U i:x i <ˆx i Simple version (implemented): move ˆx to closest vertex of box (ˆx i {x L i, x U i }) g(ˆx) + d L i (x i ˆx i ) + di U (x i ˆx i ) g(x) (x i ˆx i ) g(x) i:ˆx i =x L i i:ˆx i =x U i fast, exact, weak, improves by branching Extending SCIP for solving mixed-integer nonlinear programs 12 / 31

38 Nonlinear Constraints: Separation by Quadratic Cuts g(x) 0, g C 2 ([x L, x U ], R), ˆx such that g(ˆx) > 0 Extending SCIP for solving mixed-integer nonlinear programs 13 / 31

39 Nonlinear Constraints: Separation by Quadratic Cuts g(x) 0, g C 2 ([x L, x U ], R), ˆx such that g(ˆx) > 0 Aim: Find q(x) = x T Ax + b T x + c that solves min A,b,c x [x L,x U ] g(x) q(x)dx such that q(x) g(x), for all x [x L, x U ], q(ˆx) = g(ˆx). Extending SCIP for solving mixed-integer nonlinear programs 13 / 31

40 Nonlinear Constraints: Separation by Quadratic Cuts g(x) 0, g C 2 ([x L, x U ], R), ˆx such that g(ˆx) > 0 Approach: Compute by solving min A,b,c q(x) = x T Ax + b T x + c g(x) q(x)dx x S such that q(x) g(x), for all x S, for a sample set S [ x L, x U]. Improve sample set adaptively. q(ˆx) = g(ˆx), Extending SCIP for solving mixed-integer nonlinear programs 13 / 31

41 Computation of Quadratic Cuts initial choice: S = vert([x L, x U ]) {random points} f(x) x^ Extending SCIP for solving mixed-integer nonlinear programs 14 / 31

42 Computation of Quadratic Cuts initial choice: S = vert([x L, x U ]) {random points} f(x) q(x) x^ Extending SCIP for solving mixed-integer nonlinear programs 14 / 31

43 Computation of Quadratic Cuts for x S with g(x) = q(x), maximize the error q(x) g(x) x f(x) q(x) x^ x* Extending SCIP for solving mixed-integer nonlinear programs 14 / 31

44 Computation of Quadratic Cuts for x S with g(x) = q(x), maximize the error q(x) g(x) x if q(x ) g(x ) > δ tol, add x to S and recompute q(x) f(x) q(x) x^ x* Extending SCIP for solving mixed-integer nonlinear programs 14 / 31

45 Computation of Quadratic Cuts for x S with g(x) = q(x), maximize the error q(x) g(x) δ max f(x) x^ δ max q(x) Extending SCIP for solving mixed-integer nonlinear programs 14 / 31

46 Computation of Quadratic Cuts for x S with g(x) = q(x), maximize the error q(x) g(x) δ max if δ max < δ tol, lower q(x) by δ max f(x) x^ δ max q(x) Extending SCIP for solving mixed-integer nonlinear programs 14 / 31

47 Separation by Quadratic Cuts (cont.) heuristic: cut is only probably valid due to sampling approach Extending SCIP for solving mixed-integer nonlinear programs 15 / 31

48 Separation by Quadratic Cuts (cont.) heuristic: cut is only probably valid due to sampling approach expensive to compute: solve several NLPs and LPs Extending SCIP for solving mixed-integer nonlinear programs 15 / 31

49 Separation by Quadratic Cuts (cont.) heuristic: cut is only probably valid due to sampling approach expensive to compute: solve several NLPs and LPs tight: derived directly from function values Extending SCIP for solving mixed-integer nonlinear programs 15 / 31

50 Separation by Quadratic Cuts (cont.) heuristic: cut is only probably valid due to sampling approach expensive to compute: solve several NLPs and LPs tight: derived directly from function values works in low dimensions Extending SCIP for solving mixed-integer nonlinear programs 15 / 31

51 Nonlinear Constraints: Domain Propagation Represent algebraic structure of problem in a directed acyclic graph (DAG): nodes: variables, operations, constraints arcs: flow of computation Extending SCIP for solving mixed-integer nonlinear programs 16 / 31

52 Nonlinear Constraints: Domain Propagation Represent algebraic structure of problem in a directed acyclic graph (DAG): nodes: variables, operations, constraints arcs: flow of computation Example: x + 2 xy + 2 y [, 7] x 2 y 2xy + 3 y [0, 2] x, y [1, 16] [1,16] [1,16] x y [,7] [0,2] 3 Extending SCIP for solving mixed-integer nonlinear programs 16 / 31

53 Nonlinear Constraints: Domain Propagation Represent algebraic structure of problem in a directed acyclic graph (DAG): nodes: variables, operations, constraints arcs: flow of computation Example: x + 2 xy + 2 y [, 7] x 2 y 2xy + 3 y [0, 2] x, y [1, 16] Forward propagation: compute bounds on intermediate nodes (top-down) [1,16] [1,16] x y [1,4] [1,256] [1,256] [1,4] 2 [1,16] [1,1024] [5,7] [0,2] Extending SCIP for solving mixed-integer nonlinear programs 16 / 31

54 Nonlinear Constraints: Domain Propagation Represent algebraic structure of problem in a directed acyclic graph (DAG): nodes: variables, operations, constraints arcs: flow of computation Example: x + 2 xy + 2 y [, 7] x 2 y 2xy + 3 y [0, 2] x, y [1, 16] Forward propagation: compute bounds on intermediate nodes (top-down) Backward propagation: reduce bounds by reverse operations (bottom-up) [1,9] [1,16] x y [1,3] [1,256] [1,16] [1,4] 2 [1,4] [1,511] [5,7] [0,2] Extending SCIP for solving mixed-integer nonlinear programs 16 / 31

55 Nonlinear Constraints: Domain Propagation Represent algebraic structure of problem in a directed acyclic graph (DAG): nodes: variables, operations, constraints arcs: flow of computation Example: x + 2 xy + 2 y [, 7] x 2 y 2xy + 3 y [0, 2] x, y [1, 16] [1,9] [1,16] x y [1,3] [1,256] [1,16] [1,4] 2 Forward propagation: compute bounds on intermediate nodes (top-down) Backward propagation: reduce 2 2 bounds by reverse operations + + (bottom-up) [5,7] implemented in Couenne [Belotti et.al.] (open-source) [1,4] [1,511] Extending SCIP for solving mixed-integer nonlinear programs 16 / 31 2 [0,2] 3

56 Overview 1 Introduction 2 Quadratic Constraints 3 Nonlinear Constraints (MINLP) 4 Heuristic 5 Preliminary Numerical Results 6 Latest Work Extending SCIP for solving mixed-integer nonlinear programs 17 / 31

57 Heuristics nothing fancy yet... solve (locally) MINLP with discrete variables fixed NLP min Extending SCIP for solving mixed-integer nonlinear programs 18 / 31

58 Heuristics nothing fancy yet... solve (locally) MINLP with discrete variables fixed NLP get fixation of discrete variables from integral solution of LP relaxation in nodes min Extending SCIP for solving mixed-integer nonlinear programs 18 / 31

59 Heuristics nothing fancy yet... solve (locally) MINLP with discrete variables fixed NLP get fixation of discrete variables from integral solution of LP relaxation in nodes integral solution found by one of SCIPs MIP heuristics min Extending SCIP for solving mixed-integer nonlinear programs 18 / 31

60 Overview 1 Introduction 2 Quadratic Constraints 3 Nonlinear Constraints (MINLP) 4 Heuristic 5 Preliminary Numerical Results 6 Latest Work Extending SCIP for solving mixed-integer nonlinear programs 19 / 31

61 Implementation Implementation using some other software packages: reading problem via GAMS interface Gradients and Hessians via CppAD LP solver: CPLEX 11.2 local search heuristic by GAMS NLP solver CONOPT 3 quadratic estimators by LaGO subproblems (max error g(x) q(x)) by Ipopt domain propagation on DAGs by Couenne eigenvalues via Lapack projects.coin-or.org/lago projects.coin-or.org/ipopt projects.coin-or.org/couenne Extending SCIP for solving mixed-integer nonlinear programs 20 / 31

62 Implementation Implementation using some other software packages: reading problem via GAMS interface Gradients and Hessians via CppAD LP solver: CPLEX 11.2 local search heuristic by GAMS NLP solver CONOPT 3 quadratic estimators by LaGO subproblems (max error g(x) q(x)) by Ipopt domain propagation on DAGs by Couenne eigenvalues via Lapack projects.coin-or.org/lago projects.coin-or.org/ipopt projects.coin-or.org/couenne -lscip.linux.x86.gnu.opt -lobjscip.linux.x86.gnu.opt -llpicpx.linux.x86.gnu.opt -llago -llagointerfaceos -losicpx -losi -lcoinutils -lipopt -lcplex -los -lcouenne -lbonmin -lcbc -lcbcsolver -losiclp -lclp -lcgl -losi -lcoinutils -llagointerfacegams libsmag.a gclib.a libg2d.a clicelib.a libf90pallib.a iolib.a -lpthread -lm -ldl -lz -lreadline -lncurses -lgfortranbegin -lgfortran -lm -lgcc_s Extending SCIP for solving mixed-integer nonlinear programs 20 / 31

63 Testset Model instances: 49 MINLPs from [Belotti et.al. 2008] in average: 200 variables (105 discrete), 151 constraints 16 convex, 18 nonconvex quadratic, 15 nonconvex nonquadratic cecil_13 classical_40_0 classical_40_1 clay0203h clay0204h csched1 csched2 c-schedule4fur7feed du-opt5 du-opt eniplac enpro48pb enpro56pb ex1233 ex1243 ex1244 ex1252 fo7 ibell3a ibienst1 imisc07 iran8x32 lop97icx m6 multistage no7_ar2_1 no7_ar3_1 no7_ar4_1 nous1 nous2 nvs19 nvs23 o7_2 par72 robust_30_0 robust_30_1 shortfall_30_0 shortfall_30_1 space25a space25 stockcycle synheat synheatmod tln12 tln5 tln6 tln7 tls5 tls6 Time limit: 1 hour Gap tolerance: 0.01% Extending SCIP for solving mixed-integer nonlinear programs 21 / 31

64 Results on convex models separation by linearization of constraint functions (1st order Taylor) instance var discr nlnz time[s] gap classical_40_ < 0.01% classical_40_ < 0.01% clay0203h < 0.01% clay0204h < 0.01% du-opt < 0.01% du-opt < 0.01% fo < 0.01% ibell3a < 0.01% m < 0.01% no7_ar2_ < 0.01% no7_ar3_ < 0.01% no7_ar4_ < 0.01% o7_ < 0.01% stockcycle < 0.01% tls % tls % Extending SCIP for solving mixed-integer nonlinear programs 22 / 31

65 Results on nonconvex quadratic models separation by McCormick underestimators instance var discr nlnz time[s] gap ibienst < 0.01% imisc < 0.01% iran8x < 0.01% lop97icx % nous % nous < 0.01% nvs < 0.01% nvs < 0.01% robust_30_ % robust_30_ % shortfall_30_ % shortfall_30_ % space space25a % tln < 0.01% tln % tln % tln % Extending SCIP for solving mixed-integer nonlinear programs 23 / 31

66 Results on nonconvex nonquadratic models separation by interval gradient cuts and quadratic cuts instance var discr nlnz time[s] gap cecil_ < 0.01% csched < 0.01% csched < 0.01% c-sched eniplac < 0.01% enpro48pb < 0.01% enpro56pb < 0.01% ex < 0.01% ex < 0.01% ex < 0.01% ex < 0.01% multistage % par synheat < 0.01% synheatmod % Extending SCIP for solving mixed-integer nonlinear programs 24 / 31

67 Comparison with BARON and LindoGlobal Metric: Solved = gap < 0.01% Extending SCIP for solving mixed-integer nonlinear programs 25 / 31

68 Overview 1 Introduction 2 Quadratic Constraints 3 Nonlinear Constraints (MINLP) 4 Heuristic 5 Preliminary Numerical Results 6 Latest Work Extending SCIP for solving mixed-integer nonlinear programs 26 / 31

69 { X := (x, y) R n R + : Second Order Cone Constraints } n (α i x i ) 2 (βy) 2 i=1 Extending SCIP for solving mixed-integer nonlinear programs 27 / 31

70 X := { (x, y) R n R + : Second Order Cone Constraints } n (α i x i ) 2 (βy) 2 i=1 Ben-Tal and Nemirovski (2001): X can be linearly outer-approximated within arbitrary precision For n = 2 introduce variables a j, b j, j = 0,..., N, and constraints π π a 0 α 1x 1, a j = cos a 2 j+1 j 1 + sin b 2 j+1 j 1, a N βy π π π b 0 α 2 x 2, b j sin a 2 j+1 j 1 + cos b 2 j+1 j 1, b N tan a 2 N+1 N. Extending SCIP for solving mixed-integer nonlinear programs 27 / 31

71 X := { (x, y) R n R + : Second Order Cone Constraints } n (α i x i ) 2 (βy) 2 i=1 Ben-Tal and Nemirovski (2001): X can be linearly outer-approximated within arbitrary precision For n = 2 introduce variables a j, b j, j = 0,..., N, and constraints π π a 0 α 1x 1, a j = cos a 2 j+1 j 1 + sin b 2 j+1 j 1, a N βy π π π b 0 α 2 x 2, b j sin a 2 j+1 j 1 + cos b 2 j+1 j 1, b N tan a 2 N+1 N. N determines approximation error δ(n) = 1 cos ( π 2 N+1 ) 1 = O(4 N ) Extending SCIP for solving mixed-integer nonlinear programs 27 / 31

72 X := { (x, y) R n R + : Second Order Cone Constraints } n (α i x i ) 2 (βy) 2 i=1 Ben-Tal and Nemirovski (2001): X can be linearly outer-approximated within arbitrary precision For n = 2 introduce variables a j, b j, j = 0,..., N, and constraints π π a 0 α 1x 1, a j = cos a 2 j+1 j 1 + sin b 2 j+1 j 1, a N βy π π π b 0 α 2 x 2, b j sin a 2 j+1 j 1 + cos b 2 j+1 j 1, b N tan a 2 N+1 N. 1 N determines approximation error δ(n) = cos ( ) π 1 = O(4 N ) 2 N+1 For n > 2, reformulate as set of SOC constraints: n/2 (α i x i ) 2 z1 2, i=1 n i= n/2 +1 (α i x i ) 2 z 2 2, z z 2 2 (βy) 2 Extending SCIP for solving mixed-integer nonlinear programs 27 / 31

73 Second Order Cone (SOC) Constraints: First Results if quadratic constraint is SOC, add linear outer-approximation variables and constraints, but keep also original constraint n = 30 robust30-0 robust30-1 shortfall30-0 shortfall30-1 Extending SCIP for solving mixed-integer nonlinear programs 28 / 31

74 Second Order Cone (SOC) Constraints: First Results if quadratic constraint is SOC, add linear outer-approximation variables and constraints, but keep also original constraint n = 30 robust30-0 robust30-1 shortfall30-0 shortfall30-1 N = 0 root gap 52% 51% 12% 14% final gap 48% 48% 4% 6% time 1h 1h 1h 1h nodes Extending SCIP for solving mixed-integer nonlinear programs 28 / 31

75 Second Order Cone (SOC) Constraints: First Results if quadratic constraint is SOC, add linear outer-approximation variables and constraints, but keep also original constraint n = 30 robust30-0 robust30-1 shortfall30-0 shortfall30-1 N = 0 root gap 52% 51% 12% 14% final gap 48% 48% 4% 6% time 1h 1h 1h 1h nodes N = 3 root gap 5.2% 2.8% 0.6% 0.8% time 27s 17s 472s 781s nodes Extending SCIP for solving mixed-integer nonlinear programs 28 / 31

76 Second Order Cone (SOC) Constraints: First Results if quadratic constraint is SOC, add linear outer-approximation variables and constraints, but keep also original constraint n = 30 robust30-0 robust30-1 shortfall30-0 shortfall30-1 N = 0 root gap 52% 51% 12% 14% final gap 48% 48% 4% 6% time 1h 1h 1h 1h nodes N = 3 root gap 5.2% 2.8% 0.6% 0.8% time 27s 17s 472s 781s nodes N = 5 root gap 0.2% 0.8% 0% 0.4% time 3s 3s 3s 18s nodes Extending SCIP for solving mixed-integer nonlinear programs 28 / 31

77 Second Order Cone (SOC) Constraints: First Results if quadratic constraint is SOC, add linear outer-approximation variables and constraints, but keep also original constraint n = 30 robust30-0 robust30-1 shortfall30-0 shortfall30-1 N = 0 root gap 52% 51% 12% 14% final gap 48% 48% 4% 6% time 1h 1h 1h 1h nodes N = 3 root gap 5.2% 2.8% 0.6% 0.8% time 27s 17s 472s 781s nodes N = 5 root gap 0.2% 0.8% 0% 0.4% time 3s 3s 3s 18s nodes N = 10 root gap 0% 0.6% 0% 0.3% time 7s 8s 6s 39s nodes Extending SCIP for solving mixed-integer nonlinear programs 28 / 31

78 Second Order Cone (SOC) Constraints: First Results if quadratic constraint is SOC, add linear outer-approximation variables and constraints, but keep also original constraint n = 30 robust30-0 robust30-1 shortfall30-0 shortfall30-1 N = 0 root gap 52% 51% 12% 14% final gap 48% 48% 4% 6% time 1h 1h 1h 1h nodes N = 3 root gap 5.2% 2.8% 0.6% 0.8% time 27s 17s 472s 781s nodes N = 5 root gap 0.2% 0.8% 0% 0.4% time 3s 3s 3s 18s nodes N = 10 root gap 0% 0.6% 0% 0.3% time 7s 8s 6s 39s nodes N = 20 root gap 0% 0.6% 0% 0.3% time 166s 71s 15s 675s nodes Extending SCIP for solving mixed-integer nonlinear programs 28 / 31

79 End Thank you! Extending SCIP for solving mixed-integer nonlinear programs 29 / 31

80 Overview 7 Branching Rule Extending SCIP for solving mixed-integer nonlinear programs 30 / 31

81 Branching Variable Selection Which variable to select for branching? if LP relaxation solution is fractional, choose integer variable (SCIP default: reliability branching on pseudo cost values) Extending SCIP for solving mixed-integer nonlinear programs 31 / 31

82 Branching Variable Selection Which variable to select for branching? if LP relaxation solution is fractional, choose integer variable (SCIP default: reliability branching on pseudo cost values) selection of continuous variable: look at Belotti et.al.: Branching and bounds tightening techniques for non-convex MINLP Extending SCIP for solving mixed-integer nonlinear programs 31 / 31

83 Branching Variable Selection Which variable to select for branching? if LP relaxation solution is fractional, choose integer variable (SCIP default: reliability branching on pseudo cost values) selection of continuous variable: look at Belotti et.al.: Branching and bounds tightening techniques for non-convex MINLP search for pareto-optimal branching rule w.r.t. br-plain performance and ease of implementation Extending SCIP for solving mixed-integer nonlinear programs 31 / 31

84 Branching Variable Selection Which variable to select for branching? if LP relaxation solution is fractional, choose integer variable (SCIP default: reliability branching on pseudo cost values) selection of continuous variable: look at Belotti et.al.: Branching and bounds tightening techniques for non-convex MINLP search for pareto-optimal branching rule w.r.t. br-plain performance and ease of implementation for constraint h i (x) u violated in ˆx, let φ i := (scaled) gap between h i (ˆx) and its best linear underestimator in ˆx Extending SCIP for solving mixed-integer nonlinear programs 31 / 31

85 Branching Variable Selection Which variable to select for branching? if LP relaxation solution is fractional, choose integer variable (SCIP default: reliability branching on pseudo cost values) selection of continuous variable: look at Belotti et.al.: Branching and bounds tightening techniques for non-convex MINLP search for pareto-optimal branching rule w.r.t. br-plain performance and ease of implementation for constraint h i (x) u violated in ˆx, let φ i := (scaled) gap between h i (ˆx) and its best linear underestimator in ˆx select variable x j with maximal variable infeasibility 0.1 x j in h i (x) φ i max φ i x j in h i (x) min x j in h i (x) φ i Extending SCIP for solving mixed-integer nonlinear programs 31 / 31

Extending a CIP framework for solving mixed integer nonlinear programs

Extending a CIP framework for solving mixed integer nonlinear programs Extending a CIP framework for solving mixed integer nonlinear programs Stefan Vigerske joint with T. Berthold, P. Belotti, T. Gellermann, A. Gleixner, S. Heinz, T. Koch, M. Pfetsch DFG Research Center

More information

Solving nonconvex MINLP by quadratic approximation

Solving nonconvex MINLP by quadratic approximation Solving nonconvex MINLP by quadratic approximation Stefan Vigerske DFG Research Center MATHEON Mathematics for key technologies 21/11/2008 IMA Hot Topics Workshop: Mixed-Integer Nonlinear Optimization

More information

Introduction to Constraint Integer Programming

Introduction to Constraint Integer Programming Introduction to Constraint Integer Programming Ambros M. Gleixner Zuse Institute Berlin MATHEON Berlin Mathematical School 5th Porto Meeting on Mathematics for Industry, April 1011, 2014, Porto 1 ZIB Fast

More information

Solving Mixed-Integer Nonlinear Programs

Solving Mixed-Integer Nonlinear Programs Solving Mixed-Integer Nonlinear Programs (with SCIP) Ambros M. Gleixner Zuse Institute Berlin MATHEON Berlin Mathematical School 5th Porto Meeting on Mathematics for Industry, April 10 11, 2014, Porto

More information

Analyzing the computational impact of individual MINLP solver components

Analyzing the computational impact of individual MINLP solver components Analyzing the computational impact of individual MINLP solver components Ambros M. Gleixner joint work with Stefan Vigerske Zuse Institute Berlin MATHEON Berlin Mathematical School MINLP 2014, June 4,

More information

Heuristics for nonconvex MINLP

Heuristics for nonconvex MINLP Heuristics for nonconvex MINLP Pietro Belotti, Timo Berthold FICO, Xpress Optimization Team, Birmingham, UK pietrobelotti@fico.com 18th Combinatorial Optimization Workshop, Aussois, 9 Jan 2014 ======This

More information

From structures to heuristics to global solvers

From structures to heuristics to global solvers From structures to heuristics to global solvers Timo Berthold Zuse Institute Berlin DFG Research Center MATHEON Mathematics for key technologies OR2013, 04/Sep/13, Rotterdam Outline From structures to

More information

Cutting Plane Separators in SCIP

Cutting Plane Separators in SCIP Cutting Plane Separators in SCIP Kati Wolter Zuse Institute Berlin DFG Research Center MATHEON Mathematics for key technologies 1 / 36 General Cutting Plane Method MIP min{c T x : x X MIP }, X MIP := {x

More information

Cutting Planes in SCIP

Cutting Planes in SCIP Cutting Planes in SCIP Kati Wolter Zuse-Institute Berlin Department Optimization Berlin, 6th June 2007 Outline 1 Cutting Planes in SCIP 2 Cutting Planes for the 0-1 Knapsack Problem 2.1 Cover Cuts 2.2

More information

A Branch-and-Refine Method for Nonconvex Mixed-Integer Optimization

A Branch-and-Refine Method for Nonconvex Mixed-Integer Optimization A Branch-and-Refine Method for Nonconvex Mixed-Integer Optimization Sven Leyffer 2 Annick Sartenaer 1 Emilie Wanufelle 1 1 University of Namur, Belgium 2 Argonne National Laboratory, USA IMA Workshop,

More information

Software for Integer and Nonlinear Optimization

Software for Integer and Nonlinear Optimization Software for Integer and Nonlinear Optimization Sven Leyffer, leyffer@mcs.anl.gov Mathematics & Computer Science Division Argonne National Laboratory Roger Fletcher & Jeff Linderoth Advanced Methods and

More information

MINLP: Theory, Algorithms, Applications: Lecture 3, Basics of Algorothms

MINLP: Theory, Algorithms, Applications: Lecture 3, Basics of Algorothms MINLP: Theory, Algorithms, Applications: Lecture 3, Basics of Algorothms Jeff Linderoth Industrial and Systems Engineering University of Wisconsin-Madison Jonas Schweiger Friedrich-Alexander-Universität

More information

GLOBAL OPTIMIZATION WITH GAMS/BARON

GLOBAL OPTIMIZATION WITH GAMS/BARON GLOBAL OPTIMIZATION WITH GAMS/BARON Nick Sahinidis Chemical and Biomolecular Engineering University of Illinois at Urbana Mohit Tawarmalani Krannert School of Management Purdue University MIXED-INTEGER

More information

Mixed Integer Non Linear Programming

Mixed Integer Non Linear Programming Mixed Integer Non Linear Programming Claudia D Ambrosio CNRS Research Scientist CNRS & LIX, École Polytechnique MPRO PMA 2016-2017 Outline What is a MINLP? Dealing with nonconvexities Global Optimization

More information

Some Recent Advances in Mixed-Integer Nonlinear Programming

Some Recent Advances in Mixed-Integer Nonlinear Programming Some Recent Advances in Mixed-Integer Nonlinear Programming Andreas Wächter IBM T.J. Watson Research Center Yorktown Heights, New York andreasw@us.ibm.com SIAM Conference on Optimization 2008 Boston, MA

More information

Comparing MIQCP solvers to a specialised algorithm for mine production scheduling

Comparing MIQCP solvers to a specialised algorithm for mine production scheduling Konrad-Zuse-Zentrum für Informationstechnik Berlin Takustraße 7 D-14195 Berlin-Dahlem Germany ANDREAS BLEY AMBROS M. GLEIXNER THORSTEN KOCH STEFAN VIGERSKE Comparing MIQCP solvers to a specialised algorithm

More information

Development of the new MINLP Solver Decogo using SCIP - Status Report

Development of the new MINLP Solver Decogo using SCIP - Status Report Development of the new MINLP Solver Decogo using SCIP - Status Report Pavlo Muts with Norman Breitfeld, Vitali Gintner, Ivo Nowak SCIP Workshop 2018, Aachen Table of contents 1. Introduction 2. Automatic

More information

Computational Mixed-Integer Programming

Computational Mixed-Integer Programming Computational Mixed-Integer Programming Ambros Gleixner and the SCIP team Zuse Institute Berlin gleixner@zib.de SCIP Optimization Suite http://scip.zib.de Theory and Practice of Satisfiability Solving

More information

Mixed-Integer Nonlinear Programming

Mixed-Integer Nonlinear Programming Mixed-Integer Nonlinear Programming Claudia D Ambrosio CNRS researcher LIX, École Polytechnique, France pictures taken from slides by Leo Liberti MPRO PMA 2016-2017 Motivating Applications Nonlinear Knapsack

More information

Solving Box-Constrained Nonconvex Quadratic Programs

Solving Box-Constrained Nonconvex Quadratic Programs Noname manuscript No. (will be inserted by the editor) Solving Box-Constrained Nonconvex Quadratic Programs Pierre Bonami Oktay Günlük Jeff Linderoth June 13, 2016 Abstract We present effective computational

More information

IBM Research Report. Branching and Bounds Tightening Techniques for Non-Convex MINLP

IBM Research Report. Branching and Bounds Tightening Techniques for Non-Convex MINLP RC24620 (W0808-031) August 13, 2008 Mathematics IBM Research Report Branching and Bounds Tightening Techniques for Non-Convex MINLP Pietro Belotti 1, Jon Lee 2, Leo Liberti 3, François Margot 1, Andreas

More information

Rounding-based heuristics for nonconvex MINLPs

Rounding-based heuristics for nonconvex MINLPs Mathematical Programming Computation manuscript No. (will be inserted by the editor) Rounding-based heuristics for nonconvex MINLPs Giacomo Nannicini Pietro Belotti March 30, 2011 Abstract We propose two

More information

Outline. 1 Introduction. 2 Modeling and Applications. 3 Algorithms Convex. 4 Algorithms Nonconvex. 5 The Future of MINLP? 2 / 126

Outline. 1 Introduction. 2 Modeling and Applications. 3 Algorithms Convex. 4 Algorithms Nonconvex. 5 The Future of MINLP? 2 / 126 Outline 1 Introduction 2 3 Algorithms Convex 4 Algorithms Nonconvex 5 The Future of MINLP? 2 / 126 Larry Leading Off 3 / 126 Introduction Outline 1 Introduction Software Tools and Online Resources Basic

More information

Mixed Integer Programming Solvers: from Where to Where. Andrea Lodi University of Bologna, Italy

Mixed Integer Programming Solvers: from Where to Where. Andrea Lodi University of Bologna, Italy Mixed Integer Programming Solvers: from Where to Where Andrea Lodi University of Bologna, Italy andrea.lodi@unibo.it November 30, 2011 @ Explanatory Workshop on Locational Analysis, Sevilla A. Lodi, MIP

More information

Multiperiod Blend Scheduling Problem

Multiperiod Blend Scheduling Problem ExxonMobil Multiperiod Blend Scheduling Problem Juan Pablo Ruiz Ignacio E. Grossmann Department of Chemical Engineering Center for Advanced Process Decision-making University Pittsburgh, PA 15213 1 Motivation

More information

Optimal Looping of Pipelines in Gas Networks

Optimal Looping of Pipelines in Gas Networks Zuse Institute Berlin Takustr. 7 14195 Berlin Germany RALF LENZ AND ROBERT SCHWARZ Optimal Looping of Pipelines in Gas Networks ZIB Report 16-67 (December 2016) Zuse Institute Berlin Takustr. 7 14195 Berlin

More information

A Fast Heuristic for GO and MINLP

A Fast Heuristic for GO and MINLP A Fast Heuristic for GO and MINLP John W. Chinneck, M. Shafique, Systems and Computer Engineering Carleton University, Ottawa, Canada Introduction Goal: Find a good quality GO/MINLP solution quickly. Trade

More information

Development of an algorithm for solving mixed integer and nonconvex problems arising in electrical supply networks

Development of an algorithm for solving mixed integer and nonconvex problems arising in electrical supply networks Development of an algorithm for solving mixed integer and nonconvex problems arising in electrical supply networks E. Wanufelle 1 S. Leyffer 2 A. Sartenaer 1 Ph. Toint 1 1 FUNDP, University of Namur 2

More information

Multiobjective Mixed-Integer Stackelberg Games

Multiobjective Mixed-Integer Stackelberg Games Solving the Multiobjective Mixed-Integer SCOTT DENEGRE TED RALPHS ISE Department COR@L Lab Lehigh University tkralphs@lehigh.edu EURO XXI, Reykjavic, Iceland July 3, 2006 Outline Solving the 1 General

More information

The Surprisingly Complicated Case of [Convex] Quadratic Optimization

The Surprisingly Complicated Case of [Convex] Quadratic Optimization The Surprisingly Complicated Case of [Convex] Quadratic Optimization Robert Fourer 4er@ampl.com AMPL Optimization Inc. www.ampl.com +1 773-336-AMPL U.S.-Mexico Workshop on Optimization and Its Applications

More information

Implementation of an αbb-type underestimator in the SGO-algorithm

Implementation of an αbb-type underestimator in the SGO-algorithm Implementation of an αbb-type underestimator in the SGO-algorithm Process Design & Systems Engineering November 3, 2010 Refining without branching Could the αbb underestimator be used without an explicit

More information

Introduction to Integer Linear Programming

Introduction to Integer Linear Programming Lecture 7/12/2006 p. 1/30 Introduction to Integer Linear Programming Leo Liberti, Ruslan Sadykov LIX, École Polytechnique liberti@lix.polytechnique.fr sadykov@lix.polytechnique.fr Lecture 7/12/2006 p.

More information

An Integrated Approach to Truss Structure Design

An Integrated Approach to Truss Structure Design Slide 1 An Integrated Approach to Truss Structure Design J. N. Hooker Tallys Yunes CPAIOR Workshop on Hybrid Methods for Nonlinear Combinatorial Problems Bologna, June 2010 How to Solve Nonlinear Combinatorial

More information

Basic notions of Mixed Integer Non-Linear Programming

Basic notions of Mixed Integer Non-Linear Programming Basic notions of Mixed Integer Non-Linear Programming Claudia D Ambrosio CNRS & LIX, École Polytechnique 5th Porto Meeting on Mathematics for Industry, April 10, 2014 C. D Ambrosio (CNRS) April 10, 2014

More information

A Review and Comparison of Solvers for Convex MINLP

A Review and Comparison of Solvers for Convex MINLP A Review and Comparison of Solvers for Convex MINLP Jan Kronqvist a, David E. Bernal b, Andreas Lundell c, and Ignacio E. Grossmann b a Process Design and Systems Engineering, Åbo Akademi University, Åbo,

More information

A COMPUTATIONAL COMPARISON OF SYMMETRY HANDLING METHODS FOR MIXED INTEGER PROGRAMS

A COMPUTATIONAL COMPARISON OF SYMMETRY HANDLING METHODS FOR MIXED INTEGER PROGRAMS A COMPUTATIONAL COMPARISON OF SYMMETRY HANDLING METHODS FOR MIXED INTEGER PROGRAMS MARC E. PFETSCH AND THOMAS REHN Abstract. The handling of symmetries in mixed integer programs in order to speed up the

More information

Structured Problems and Algorithms

Structured Problems and Algorithms Integer and quadratic optimization problems Dept. of Engg. and Comp. Sci., Univ. of Cal., Davis Aug. 13, 2010 Table of contents Outline 1 2 3 Benefits of Structured Problems Optimization problems may become

More information

Indicator Constraints in Mixed-Integer Programming

Indicator Constraints in Mixed-Integer Programming Indicator Constraints in Mixed-Integer Programming Andrea Lodi University of Bologna, Italy - andrea.lodi@unibo.it Amaya Nogales-Gómez, Universidad de Sevilla, Spain Pietro Belotti, FICO, UK Matteo Fischetti,

More information

Feasibility Pump for Mixed Integer Nonlinear Programs 1

Feasibility Pump for Mixed Integer Nonlinear Programs 1 Feasibility Pump for Mixed Integer Nonlinear Programs 1 Presenter: 1 by Pierre Bonami, Gerard Cornuejols, Andrea Lodi and Francois Margot Mixed Integer Linear or Nonlinear Programs (MILP/MINLP) Optimize

More information

Heuristics and Upper Bounds for a Pooling Problem with Cubic Constraints

Heuristics and Upper Bounds for a Pooling Problem with Cubic Constraints Heuristics and Upper Bounds for a Pooling Problem with Cubic Constraints Matthew J. Real, Shabbir Ahmed, Helder Inàcio and Kevin Norwood School of Chemical & Biomolecular Engineering 311 Ferst Drive, N.W.

More information

SCIP-Jack: A Solver for Steiner Tree Problems in Graphs and their Relatives

SCIP-Jack: A Solver for Steiner Tree Problems in Graphs and their Relatives SCIP-Jack: A Solver for Steiner Tree Problems in Graphs and their Relatives Thorsten Koch Daniel Rehfeldt Zuse Institute Berlin Technical University Berlin Joint Work with Gerald Gamrath Stephen Maher

More information

Advances in Bayesian Network Learning using Integer Programming

Advances in Bayesian Network Learning using Integer Programming Advances in Bayesian Network Learning using Integer Programming Mark Bartlett and James Cussens UAI-13, 2013-07-12 Supported by the UK Medical Research Council (Project Grant G1002312) Mark Bartlett and

More information

Integer Programming for Bayesian Network Structure Learning

Integer Programming for Bayesian Network Structure Learning Integer Programming for Bayesian Network Structure Learning James Cussens Prague, 2013-09-02 Supported by the UK Medical Research Council (Project Grant G1002312) James Cussens IP for BNs Prague, 2013-09-02

More information

1 Column Generation and the Cutting Stock Problem

1 Column Generation and the Cutting Stock Problem 1 Column Generation and the Cutting Stock Problem In the linear programming approach to the traveling salesman problem we used the cutting plane approach. The cutting plane approach is appropriate when

More information

Improved quadratic cuts for convex mixed-integer nonlinear programs

Improved quadratic cuts for convex mixed-integer nonlinear programs Improved quadratic cuts for convex mixed-integer nonlinear programs Lijie Su a,b, Lixin Tang a*, David E. Bernal c, Ignacio E. Grossmann c a Institute of Industrial and Systems Engineering, Northeastern

More information

Numerical Optimization. Review: Unconstrained Optimization

Numerical Optimization. Review: Unconstrained Optimization Numerical Optimization Finding the best feasible solution Edward P. Gatzke Department of Chemical Engineering University of South Carolina Ed Gatzke (USC CHE ) Numerical Optimization ECHE 589, Spring 2011

More information

MVE165/MMG631 Linear and integer optimization with applications Lecture 8 Discrete optimization: theory and algorithms

MVE165/MMG631 Linear and integer optimization with applications Lecture 8 Discrete optimization: theory and algorithms MVE165/MMG631 Linear and integer optimization with applications Lecture 8 Discrete optimization: theory and algorithms Ann-Brith Strömberg 2017 04 07 Lecture 8 Linear and integer optimization with applications

More information

18 hours nodes, first feasible 3.7% gap Time: 92 days!! LP relaxation at root node: Branch and bound

18 hours nodes, first feasible 3.7% gap Time: 92 days!! LP relaxation at root node: Branch and bound The MIP Landscape 1 Example 1: LP still can be HARD SGM: Schedule Generation Model Example 157323 1: LP rows, still can 182812 be HARD columns, 6348437 nzs LP relaxation at root node: 18 hours Branch and

More information

On handling indicator constraints in mixed integer programming

On handling indicator constraints in mixed integer programming Comput Optim Appl (2016) 65:545 566 DOI 10.1007/s10589-016-9847-8 On handling indicator constraints in mixed integer programming Pietro Belotti 1 Pierre Bonami 2 Matteo Fischetti 3 Andrea Lodi 4,5 Michele

More information

Optimization Bounds from Binary Decision Diagrams

Optimization Bounds from Binary Decision Diagrams Optimization Bounds from Binary Decision Diagrams J. N. Hooker Joint work with David Bergman, André Ciré, Willem van Hoeve Carnegie Mellon University ICS 203 Binary Decision Diagrams BDDs historically

More information

Integer Programming for Bayesian Network Structure Learning

Integer Programming for Bayesian Network Structure Learning Integer Programming for Bayesian Network Structure Learning James Cussens Helsinki, 2013-04-09 James Cussens IP for BNs Helsinki, 2013-04-09 1 / 20 Linear programming The Belgian diet problem Fat Sugar

More information

Convex Quadratic Relaxations of Nonconvex Quadratically Constrained Quadratic Progams

Convex Quadratic Relaxations of Nonconvex Quadratically Constrained Quadratic Progams Convex Quadratic Relaxations of Nonconvex Quadratically Constrained Quadratic Progams John E. Mitchell, Jong-Shi Pang, and Bin Yu Original: June 10, 2011 Abstract Nonconvex quadratic constraints can be

More information

Outline. Relaxation. Outline DMP204 SCHEDULING, TIMETABLING AND ROUTING. 1. Lagrangian Relaxation. Lecture 12 Single Machine Models, Column Generation

Outline. Relaxation. Outline DMP204 SCHEDULING, TIMETABLING AND ROUTING. 1. Lagrangian Relaxation. Lecture 12 Single Machine Models, Column Generation Outline DMP204 SCHEDULING, TIMETABLING AND ROUTING 1. Lagrangian Relaxation Lecture 12 Single Machine Models, Column Generation 2. Dantzig-Wolfe Decomposition Dantzig-Wolfe Decomposition Delayed Column

More information

Feasibility Pump Heuristics for Column Generation Approaches

Feasibility Pump Heuristics for Column Generation Approaches 1 / 29 Feasibility Pump Heuristics for Column Generation Approaches Ruslan Sadykov 2 Pierre Pesneau 1,2 Francois Vanderbeck 1,2 1 University Bordeaux I 2 INRIA Bordeaux Sud-Ouest SEA 2012 Bordeaux, France,

More information

Alternative Methods for Obtaining. Optimization Bounds. AFOSR Program Review, April Carnegie Mellon University. Grant FA

Alternative Methods for Obtaining. Optimization Bounds. AFOSR Program Review, April Carnegie Mellon University. Grant FA Alternative Methods for Obtaining Optimization Bounds J. N. Hooker Carnegie Mellon University AFOSR Program Review, April 2012 Grant FA9550-11-1-0180 Integrating OR and CP/AI Early support by AFOSR First

More information

BCOL RESEARCH REPORT 07.04

BCOL RESEARCH REPORT 07.04 BCOL RESEARCH REPORT 07.04 Industrial Engineering & Operations Research University of California, Berkeley, CA 94720-1777 LIFTING FOR CONIC MIXED-INTEGER PROGRAMMING ALPER ATAMTÜRK AND VISHNU NARAYANAN

More information

Disconnecting Networks via Node Deletions

Disconnecting Networks via Node Deletions 1 / 27 Disconnecting Networks via Node Deletions Exact Interdiction Models and Algorithms Siqian Shen 1 J. Cole Smith 2 R. Goli 2 1 IOE, University of Michigan 2 ISE, University of Florida 2012 INFORMS

More information

Integer Programming. Wolfram Wiesemann. December 6, 2007

Integer Programming. Wolfram Wiesemann. December 6, 2007 Integer Programming Wolfram Wiesemann December 6, 2007 Contents of this Lecture Revision: Mixed Integer Programming Problems Branch & Bound Algorithms: The Big Picture Solving MIP s: Complete Enumeration

More information

Hot-Starting NLP Solvers

Hot-Starting NLP Solvers Hot-Starting NLP Solvers Andreas Wächter Department of Industrial Engineering and Management Sciences Northwestern University waechter@iems.northwestern.edu 204 Mixed Integer Programming Workshop Ohio

More information

23. Cutting planes and branch & bound

23. Cutting planes and branch & bound CS/ECE/ISyE 524 Introduction to Optimization Spring 207 8 23. Cutting planes and branch & bound ˆ Algorithms for solving MIPs ˆ Cutting plane methods ˆ Branch and bound methods Laurent Lessard (www.laurentlessard.com)

More information

Mixed Integer Programming (MIP) for Causal Inference and Beyond

Mixed Integer Programming (MIP) for Causal Inference and Beyond Mixed Integer Programming (MIP) for Causal Inference and Beyond Juan Pablo Vielma Massachusetts Institute of Technology Columbia Business School New York, NY, October, 2016. Traveling Salesman Problem

More information

Exploring the Numerics of Branch-and-Cut for Mixed Integer Linear Optimization

Exploring the Numerics of Branch-and-Cut for Mixed Integer Linear Optimization Zuse Institute Berlin Takustr. 7 14195 Berlin Germany MATTHIAS MILTENBERGER 1, TED RALPHS 2, DANIEL E. STEFFY 3 Exploring the Numerics of Branch-and-Cut for Mixed Integer Linear Optimization 1 Zuse Institute

More information

Integer Linear Programs

Integer Linear Programs Lecture 2: Review, Linear Programming Relaxations Today we will talk about expressing combinatorial problems as mathematical programs, specifically Integer Linear Programs (ILPs). We then see what happens

More information

Integer Programming Chapter 15

Integer Programming Chapter 15 Integer Programming Chapter 15 University of Chicago Booth School of Business Kipp Martin November 9, 2016 1 / 101 Outline Key Concepts Problem Formulation Quality Solver Options Epsilon Optimality Preprocessing

More information

Lecture 9: Dantzig-Wolfe Decomposition

Lecture 9: Dantzig-Wolfe Decomposition Lecture 9: Dantzig-Wolfe Decomposition (3 units) Outline Dantzig-Wolfe decomposition Column generation algorithm Relation to Lagrangian dual Branch-and-price method Generated assignment problem and multi-commodity

More information

On mathematical programming with indicator constraints

On mathematical programming with indicator constraints On mathematical programming with indicator constraints Andrea Lodi joint work with P. Bonami & A. Tramontani (IBM), S. Wiese (Unibo) University of Bologna, Italy École Polytechnique de Montréal, Québec,

More information

Advances in CPLEX for Mixed Integer Nonlinear Optimization

Advances in CPLEX for Mixed Integer Nonlinear Optimization Pierre Bonami and Andrea Tramontani IBM ILOG CPLEX ISMP 2015 - Pittsburgh - July 13 2015 Advances in CPLEX for Mixed Integer Nonlinear Optimization 1 2015 IBM Corporation CPLEX Optimization Studio 12.6.2

More information

Accelerating the Convergence of Stochastic Unit Commitment Problems by Using Tight and Compact MIP Formulations

Accelerating the Convergence of Stochastic Unit Commitment Problems by Using Tight and Compact MIP Formulations Accelerating the Convergence of Stochastic Unit Commitment Problems by Using Tight and Compact MIP Formulations Germán Morales-España, and Andrés Ramos Delft University of Technology, Delft, The Netherlands

More information

Using Sparsity to Design Primal Heuristics for MILPs: Two Stories

Using Sparsity to Design Primal Heuristics for MILPs: Two Stories for MILPs: Two Stories Santanu S. Dey Joint work with: Andres Iroume, Marco Molinaro, Domenico Salvagnin, Qianyi Wang MIP Workshop, 2017 Sparsity in real" Integer Programs (IPs) Real" IPs are sparse: The

More information

Presolve Reductions in Mixed Integer Programming

Presolve Reductions in Mixed Integer Programming Zuse Institute Berlin Takustr. 7 14195 Berlin Germany TOBIAS ACHTERBERG, ROBERT E. BIXBY, ZONGHAO GU, EDWARD ROTHBERG, AND DIETER WENINGER Presolve Reductions in Mixed Integer Programming This work has

More information

Enhancing Fathoming Rules in Branch-and-Bound for Biobjective Mixed-Integer Programming

Enhancing Fathoming Rules in Branch-and-Bound for Biobjective Mixed-Integer Programming Enhancing Fathoming Rules in Branch-and-Bound for Biobjective Mixed-Integer Programming Nathan Adelgren Clemson University MIP - July 21, 2014 Joint Work with: Dr. Pietro Belotti - FICO (Birmingham, UK)

More information

On Solving Aircraft Conflict Avoidance Using Deterministic Global Optimization (sbb) Codes

On Solving Aircraft Conflict Avoidance Using Deterministic Global Optimization (sbb) Codes On Solving Aircraft Conflict Avoidance Using Deterministic Global Optimization (sbb) Codes Sonia Cafieri, Frédéric Messine, Ahmed Touhami To cite this version: Sonia Cafieri, Frédéric Messine, Ahmed Touhami.

More information

Module 04 Optimization Problems KKT Conditions & Solvers

Module 04 Optimization Problems KKT Conditions & Solvers Module 04 Optimization Problems KKT Conditions & Solvers Ahmad F. Taha EE 5243: Introduction to Cyber-Physical Systems Email: ahmad.taha@utsa.edu Webpage: http://engineering.utsa.edu/ taha/index.html September

More information

Preprocessing. Complements of Operations Research. Giovanni Righini. Università degli Studi di Milano

Preprocessing. Complements of Operations Research. Giovanni Righini. Università degli Studi di Milano Preprocessing Complements of Operations Research Giovanni Righini Università degli Studi di Milano Preprocessing Computational complexity theory classifies problems. However, when we run algorithms, this

More information

{0, 1 2. }-Chvátal-Gomory Cuts. Algorithms to Separate. Konrad-Zuse-Zentrum für Informationstechnik Berlin

{0, 1 2. }-Chvátal-Gomory Cuts. Algorithms to Separate. Konrad-Zuse-Zentrum für Informationstechnik Berlin Konrad-Zuse-Zentrum für Informationstechnik Berlin Takustraße 7 D-14195 Berlin-Dahlem Germany ARIE M.C.A. KOSTER ADRIAN ZYMOLKA MANUEL KUTSCHKA Algorithms to Separate {0, 1 2 }-Chvátal-Gomory Cuts ZIB-Report

More information

Computational testing of exact separation for mixed-integer knapsack problems

Computational testing of exact separation for mixed-integer knapsack problems Computational testing of exact separation for mixed-integer knapsack problems Pasquale Avella (joint work with Maurizio Boccia and Igor Vasiliev ) DING - Università del Sannio Russian Academy of Sciences

More information

Machine learning, ALAMO, and constrained regression

Machine learning, ALAMO, and constrained regression Machine learning, ALAMO, and constrained regression Nick Sahinidis Acknowledgments: Alison Cozad, David Miller, Zach Wilson MACHINE LEARNING PROBLEM Build a model of output variables as a function of input

More information

Quick Tour of Linear Algebra and Graph Theory

Quick Tour of Linear Algebra and Graph Theory Quick Tour of Linear Algebra and Graph Theory CS224w: Social and Information Network Analysis Fall 2012 Yu Wayne Wu Based on Borja Pelato s version in Fall 2011 Matrices and Vectors Matrix: A rectangular

More information

OQNLP: a Scatter Search Multistart Approach for Solving Constrained Non- Linear Global Optimization Problems

OQNLP: a Scatter Search Multistart Approach for Solving Constrained Non- Linear Global Optimization Problems OQNLP: a Scatter Search Multistart Approach for Solving Constrained Non- Linear Global Optimization Problems Zsolt Ugray, The University of California at Riverside, MIS Dept. Leon Lasdon, The University

More information

Copositive Programming and Combinatorial Optimization

Copositive Programming and Combinatorial Optimization Copositive Programming and Combinatorial Optimization Franz Rendl http://www.math.uni-klu.ac.at Alpen-Adria-Universität Klagenfurt Austria joint work with I.M. Bomze (Wien) and F. Jarre (Düsseldorf) IMA

More information

RLT-POS: Reformulation-Linearization Technique (RLT)-based Optimization Software for Polynomial Programming Problems

RLT-POS: Reformulation-Linearization Technique (RLT)-based Optimization Software for Polynomial Programming Problems RLT-POS: Reformulation-Linearization Technique (RLT)-based Optimization Software for Polynomial Programming Problems E. Dalkiran 1 H.D. Sherali 2 1 The Department of Industrial and Systems Engineering,

More information

Chapter 4: Interpolation and Approximation. October 28, 2005

Chapter 4: Interpolation and Approximation. October 28, 2005 Chapter 4: Interpolation and Approximation October 28, 2005 Outline 1 2.4 Linear Interpolation 2 4.1 Lagrange Interpolation 3 4.2 Newton Interpolation and Divided Differences 4 4.3 Interpolation Error

More information

Convex Relaxations of Non-Convex Mixed Integer Quadratically Constrained Programs: Projected Formulations

Convex Relaxations of Non-Convex Mixed Integer Quadratically Constrained Programs: Projected Formulations Anureet Saxena Pierre Bonami Jon Lee Convex Relaxations of Non-Convex Mixed Integer Quadratically Constrained Programs: Projected Formulations November 19, 008 Abstract A common way to produce a convex

More information

Lifted Inequalities for 0 1 Mixed-Integer Bilinear Covering Sets

Lifted Inequalities for 0 1 Mixed-Integer Bilinear Covering Sets 1 2 3 Lifted Inequalities for 0 1 Mixed-Integer Bilinear Covering Sets Kwanghun Chung 1, Jean-Philippe P. Richard 2, Mohit Tawarmalani 3 March 1, 2011 4 5 6 7 8 9 Abstract In this paper, we study 0 1 mixed-integer

More information

Branch-and-Cut for Linear Programs with Overlapping SOS1 Constraints

Branch-and-Cut for Linear Programs with Overlapping SOS1 Constraints Noname manuscript No. (will be inserted by the editor) Branch-and-Cut for Linear Programs with Overlapping SOS1 Constraints Tobias Fischer Marc E. Pfetsch Received: date / Accepted: date April 1, 2015

More information

The moment-lp and moment-sos approaches

The moment-lp and moment-sos approaches The moment-lp and moment-sos approaches LAAS-CNRS and Institute of Mathematics, Toulouse, France CIRM, November 2013 Semidefinite Programming Why polynomial optimization? LP- and SDP- CERTIFICATES of POSITIVITY

More information

Functions of Several Variables

Functions of Several Variables Functions of Several Variables The Unconstrained Minimization Problem where In n dimensions the unconstrained problem is stated as f() x variables. minimize f()x x, is a scalar objective function of vector

More information

Probabilistic Graphical Models

Probabilistic Graphical Models Probabilistic Graphical Models David Sontag New York University Lecture 6, March 7, 2013 David Sontag (NYU) Graphical Models Lecture 6, March 7, 2013 1 / 25 Today s lecture 1 Dual decomposition 2 MAP inference

More information

A FRAMEWORK FOR SOLVING MIXED-INTEGER SEMIDEFINITE PROGRAMS

A FRAMEWORK FOR SOLVING MIXED-INTEGER SEMIDEFINITE PROGRAMS A FRAMEWORK FOR SOLVING MIXED-INTEGER SEMIDEFINITE PROGRAMS TRISTAN GALLY, MARC E. PFETSCH, AND STEFAN ULBRICH ABSTRACT. Mixed-integer semidefinite programs arise in many applications and several problem-specific

More information

Using Convex Nonlinear Relaxations in the Global Optimization of Nonconvex Generalized Disjunctive Programs

Using Convex Nonlinear Relaxations in the Global Optimization of Nonconvex Generalized Disjunctive Programs Using Convex Nonlinear Relaxations in the Global Optimization of Nonconvex Generalized Disjunctive Programs Juan P. Ruiz, Ignacio E. Grossmann Carnegie Mellon University - Department of Chemical Engineering

More information

Monomial-wise Optimal Separable Underestimators for Mixed-Integer Polynomial Optimization

Monomial-wise Optimal Separable Underestimators for Mixed-Integer Polynomial Optimization Monomial-wise Optimal Separable Underestimators for Mixed-Integer Polynomial Optimization Christoph Buchheim Claudia D Ambrosio Received: date / Accepted: date Abstract In this paper we introduce a new

More information

Lecture 10: Duality in Linear Programs

Lecture 10: Duality in Linear Programs 10-725/36-725: Convex Optimization Spring 2015 Lecture 10: Duality in Linear Programs Lecturer: Ryan Tibshirani Scribes: Jingkun Gao and Ying Zhang Disclaimer: These notes have not been subjected to the

More information

Integer Programming ISE 418. Lecture 8. Dr. Ted Ralphs

Integer Programming ISE 418. Lecture 8. Dr. Ted Ralphs Integer Programming ISE 418 Lecture 8 Dr. Ted Ralphs ISE 418 Lecture 8 1 Reading for This Lecture Wolsey Chapter 2 Nemhauser and Wolsey Sections II.3.1, II.3.6, II.4.1, II.4.2, II.5.4 Duality for Mixed-Integer

More information

CHAPTER 2: QUADRATIC PROGRAMMING

CHAPTER 2: QUADRATIC PROGRAMMING CHAPTER 2: QUADRATIC PROGRAMMING Overview Quadratic programming (QP) problems are characterized by objective functions that are quadratic in the design variables, and linear constraints. In this sense,

More information

Stochastic Integer Programming

Stochastic Integer Programming IE 495 Lecture 20 Stochastic Integer Programming Prof. Jeff Linderoth April 14, 2003 April 14, 2002 Stochastic Programming Lecture 20 Slide 1 Outline Stochastic Integer Programming Integer LShaped Method

More information

Optimization Problems in Gas Transportation. Lars Schewe Friedrich-Alexander-Universität Erlangen-Nürnberg CWM 3 EO 2014, Budapest

Optimization Problems in Gas Transportation. Lars Schewe Friedrich-Alexander-Universität Erlangen-Nürnberg CWM 3 EO 2014, Budapest Optimization Problems in Gas Transportation Lars Schewe Friedrich-Alexander-Universität Erlangen-Nürnberg CWM 3 EO 2014, Budapest Why do we care? Lars Schewe FAU Optimization Problems in Gas Transportation

More information

Decomposition-based Methods for Large-scale Discrete Optimization p.1

Decomposition-based Methods for Large-scale Discrete Optimization p.1 Decomposition-based Methods for Large-scale Discrete Optimization Matthew V Galati Ted K Ralphs Department of Industrial and Systems Engineering Lehigh University, Bethlehem, PA, USA Départment de Mathématiques

More information

Bilevel Integer Optimization: Theory and Algorithms

Bilevel Integer Optimization: Theory and Algorithms : Theory and Algorithms Ted Ralphs 1 Joint work with Sahar Tahernajad 1, Scott DeNegre 3, Menal Güzelsoy 2, Anahita Hassanzadeh 4 1 COR@L Lab, Department of Industrial and Systems Engineering, Lehigh University

More information

MATH 4211/6211 Optimization Basics of Optimization Problems

MATH 4211/6211 Optimization Basics of Optimization Problems MATH 4211/6211 Optimization Basics of Optimization Problems Xiaojing Ye Department of Mathematics & Statistics Georgia State University Xiaojing Ye, Math & Stat, Georgia State University 0 A standard minimization

More information

Multidisciplinary System Design Optimization (MSDO)

Multidisciplinary System Design Optimization (MSDO) Multidisciplinary System Design Optimization (MSDO) Numerical Optimization II Lecture 8 Karen Willcox 1 Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox Today s Topics Sequential

More information