Outline. 1 Introduction. 2 Modeling and Applications. 3 Algorithms Convex. 4 Algorithms Nonconvex. 5 The Future of MINLP? 2 / 126
|
|
- Audrey Gilmore
- 5 years ago
- Views:
Transcription
1 Outline 1 Introduction 2 3 Algorithms Convex 4 Algorithms Nonconvex 5 The Future of MINLP? 2 / 126
2 Larry Leading Off 3 / 126
3 Introduction Outline 1 Introduction Software Tools and Online Resources Basic Concepts 2 3 Algorithms Convex 4 Algorithms Nonconvex 5 The Future of MINLP? 4 / 126
4 Introduction Mixed-Integer Nonlinear Optimization Mixed Integer Nonlinear Program: (MINLP) z minlp = min x f (x) subject to g i (x) 0, i = 1,..., m x X, x I Z I f : R n R, g : R n R m smooth, sometimes convex functions. X R n bounded, polyhedral set, e.g. X = {x : l A T x u} I {1,..., n} subset of integer variables 5 / 126
5 Introduction Mixed-Integer Nonlinear Optimization Mixed Integer Nonlinear Program: (MINLP) z minlp = min x η subject to g i (x) 0, i = 1,..., m f (x) η x X, x I Z I f : R n R, g : R n R m smooth, sometimes convex functions. X R n bounded, polyhedral set, e.g. X = {x : l A T x u} I {1,..., n} subset of integer variables We may assume that f is a linear function. 5 / 126
6 Introduction Mixed-Integer Nonlinear Optimization Mixed Integer Nonlinear Program: (MINLP) z minlp = min x η subject to g i (x) 0, i = 1,..., m f (x) η x X, x I Z I f : R n R, g : R n R m smooth, sometimes convex functions. X R n bounded, polyhedral set, e.g. X = {x : l A T x u} I {1,..., n} subset of integer variables We may assume that f is a linear function. NP-Super-Hard Combines challenges of handling nonlinearities with combinatorial explosion of integer variables 5 / 126
7 Introduction The MINLP Family The great watershed in optimization isn t between linearity and nonlinearity, but convexity and nonconvexity. - R. Tyrrell Rockafellar If f and g are convex functions, then we have a convex MINLP If f and g are not convex, then we have a nonconvex MINLP 6 / 126
8 Introduction The MINLP Family The great watershed in optimization isn t between linearity and nonlinearity, but convexity and nonconvexity. - R. Tyrrell Rockafellar If f and g are convex functions, then we have a convex MINLP If f and g are not convex, then we have a nonconvex MINLP Optimization 101 Know the convexity properties of your instance! Algorithms for two different classes of problems are different We can solve significantly larger convex MINLP instances that nonconvex MINLP instances Algorithms designed for convex-minlp are often used as heuristics for nonconvex MINLPs 6 / 126
9 Introduction Specializations Functional Form g i (x) = A i x + b i 2 p T i x + q i g i (x) = polynomial Problem Type MISOCP MIPP g i (x) = x T Q i x + p T i + q i MIQP g i (x) = a T i x b i MILP MISOCP: Mixed Integer Second Order Cone Program Are convex-minlp MIPP: Mixed Integer Polynomial Program MIQP: Mixed Integer Quadratic Program May be convex or nonconvex Convex MIQP is a special case of MISOCP If f is convex quadratic and c is an affine mapping, then there are specialized algorithms for convex-miqp MILP: Mixed Integer Linear Program 7 / 126
10 Introduction Aside MISOCP The feasible regions of surprisingly MINLP can be expressed as MISOCP: 8 / 126
11 Introduction Aside MISOCP The feasible regions of surprisingly MINLP can be expressed as MISOCP: Hyperbolic constraints: Product of (non-negative) variables a norm 2. [ ] w T w xy, x 0, y 0 2w x + y x y 8 / 126
12 Introduction Aside MISOCP The feasible regions of surprisingly MINLP can be expressed as MISOCP: Hyperbolic constraints: Product of (non-negative) variables a norm 2. [ ] w T w xy, x 0, y 0 2w x + y x y Convex Quadratic: c i (x) = x T Q i x + p T i x + q i t Q i is positive semidefinite, so it has Cholesky factorization Q i = L T L y = Lx g(x) = y T y + p T i x + q i t {(x, t) R n+1 x T Qx + p T i x + q i t} = Proj (x,t) {(x, y, t, p) R 2n+2 y = Lx, p + q T x + r t, p y 2 } 8 / 126
13 Introduction MINLP Tree MINLP Nonconvex NLP Nonconvex MINLP Convex MINLP Convex NLP Polynomial Opt. MI- Polynomial Opt. MISOCP SOCP Nonconvex QP Nonconvex MIQP Convex MIQP Convex QP MILP LP 9 / 126
14 Introduction How Big Is It? The Leyffer-Linderoth-Luedtke (LLL) Measure of Complexity You have a problem of class X that has Y decision variables? What is the largest value of Y for which one of Sven, Jim, and Jeff would be willing to bet $50 that a state-of-the-art solver could solve the problem? 10 / 126
15 Introduction How Big Is It? The Leyffer-Linderoth-Luedtke (LLL) Measure of Complexity You have a problem of class X that has Y decision variables? What is the largest value of Y for which one of Sven, Jim, and Jeff would be willing to bet $50 that a state-of-the-art solver could solve the problem? Convex Nonconvex Problem Class (X ) # Var (Y ) Problem Class (X ) # Var (Y ) MINLP 500 MINLP 100 NLP NLP 100 MISOCP 1000 MIPP 150 SOCP 10 5 PP 150 MIQP 1000 MIQP 300 QP QP 300 MILP LP / 126
16 Introduction Software Tools and Online Resources Outline 1 Introduction Software Tools and Online Resources Basic Concepts 2 3 Algorithms Convex 4 Algorithms Nonconvex 5 The Future of MINLP? 11 / 126
17 Introduction Software Tools and Online Resources Solvers for MINLP Convex α-ecp BONMIN DICOPT FilMINT GUROBI KNITRO MILANO MINLPBB MINOPT MINOTAUR SBB Xpress MIQP Nonconvex α-bb ANTIGONE BARON COCONUT COUENNE CPLEX LGO LindoGlobal SCIP MIQP/convex MIQP Open-source and commercial solvers 12 / 126
18 Introduction Software Tools and Online Resources Optimization Modeling Languages Advantages Express optimization languages in natural algebraic form Access to automatic differentiation & presolve techniques 13 / 126
19 Introduction Software Tools and Online Resources Optimization Modeling Languages Advantages Express optimization languages in natural algebraic form Access to automatic differentiation & presolve techniques Modeling Languages for Optimization AIMMS [Bisschop and Entriken, 1993] AMPL [Fourer et al., 1993] GAMS [Brooke et al., 1992] MOSEL [Colombani and Heipcke, 2002] TomLab [Holmström and Edvall, 2004] (On top of Matlab) Pyomo [Hart et al., 2011] 13 / 126
20 Introduction Software Tools and Online Resources Online Resources MINLP Test Problem Libraries wiki.mcs.anl.gov/leyffer/index.php/macminlp minlp.org/... models & descriptions Good source of modeling tricks and practices! Online Solvers Computational Infrastructure for Operations Research 14 / 126
21 Introduction Basic Concepts Outline 1 Introduction Software Tools and Online Resources Basic Concepts 2 3 Algorithms Convex 4 Algorithms Nonconvex 5 The Future of MINLP? 15 / 126
22 Introduction Basic Concepts Simple Example x 2 (x 1 2) 2 + (x 2 + 1) x 1 x (x 1 1) 2 x x 1, x 2 5 x 2 Z Doubly NP-Hard! Feasible region is not convex for two reasons Integrality of x 2 and nonconvexity of constraint x 1 16 / 126
23 Introduction Basic Concepts Q: What to Do?! Relax! Remove some constraints. Solution of relaxed problem provide a lower bound on optimal solution value 17 / 126
24 Introduction Basic Concepts Natural Relaxation x 2 x 1 18 / 126
25 Introduction Basic Concepts Natural Relaxation x 2 Natural relaxation is to remove constraints x i Z i I But this still leaves a nonconvex NLP relaxation x 1 18 / 126
26 Introduction Basic Concepts Relax More We need a mechanism for relaxing the nonconvex constraints. x 2 x 1 19 / 126
27 Introduction Basic Concepts Relax More We need a mechanism for relaxing the nonconvex constraints. Secant underestimator for concave functions x 2 x 1 19 / 126
28 Introduction Basic Concepts Relax More We need a mechanism for relaxing the nonconvex constraints. Secant underestimator for concave functions Easy to get polyhedral outerapproximation of nonconvex constraints x 2 x 1 19 / 126
29 Introduction Basic Concepts Relax More We need a mechanism for relaxing the nonconvex constraints. Secant underestimator for concave functions Easy to get polyhedral outerapproximation of nonconvex constraints Ideally, we would like to get the convex hull of the feasible points x 2 x 1 19 / 126
30 Introduction Basic Concepts Linear Objective Is Important Here! min(x 1 1/2) 2 + (x 2 1/2) 2 x 2 s.t. x 1 {0, 1}, x 2 {0, 1} (ˆx 1, ˆx 2 ) x 1 20 / 126
31 Introduction Basic Concepts Linear Objective Is Important Here! min(x 1 1/2) 2 + (x 2 1/2) 2 x 2 s.t. x 1 {0, 1}, x 2 {0, 1} (ˆx 1, ˆx 2 ) η (x 1 1/2) 2 + (x 2 1/2) 2 η x 1 Without linear objective, optimal solution may be interior to the convex hull convexifying may do you no good! 20 / 126
32 Introduction Basic Concepts Algorithm Ingredients 1 Relaxation Used to compute a lower bound on the optimum value Obtained by enlarging feasible set; e.g. ignore constraints Should be tractable and tight 21 / 126
33 Introduction Basic Concepts Algorithm Ingredients 1 Relaxation Used to compute a lower bound on the optimum value Obtained by enlarging feasible set; e.g. ignore constraints Should be tractable and tight 2 Constraint Enforcement Exclude solutions from relaxations not feasible to MINLP Refine or tighten of relaxation Branching and Cutting 21 / 126
34 Introduction Basic Concepts Algorithm Ingredients 1 Relaxation Used to compute a lower bound on the optimum value Obtained by enlarging feasible set; e.g. ignore constraints Should be tractable and tight 2 Constraint Enforcement Exclude solutions from relaxations not feasible to MINLP Refine or tighten of relaxation Branching and Cutting 3 Upper Bounds Obtained from any feasible point; e.g. solve NLP for fixed x I. (NLP(x I )). Other heuristic techniques have been developed 21 / 126
35 Outline 1 Introduction 2 Supply Chain Water Network Design Pooling 3 Algorithms Convex 4 Algorithms Nonconvex 5 The Future of MINLP? 22 / 126
36 Models are Important! Writing good models is extremely important in MILP Probably more important for MINLP 23 / 126
37 Models are Important! Writing good models is extremely important in MILP Probably more important for MINLP More Complexity Tradeoff between physical accuracy and mathematical tractability Convex reformulations? Linearize expressions? 23 / 126
38 Models are Important! Writing good models is extremely important in MILP Probably more important for MINLP More Complexity Tradeoff between physical accuracy and mathematical tractability Convex reformulations? Linearize expressions? MINLP Modeling Preference We want strong formulations Relaxations that can be relaxed to something that closely approximates the convex hull We prefer linear over convex over nonconvex formulations 23 / 126
39 Some MINLP Applications 1 Supply Chain Convex MINLP Discrete: Fixed charges for opening facilities Nonlinear: Nonlinear transportation costs 2 Gas/Water Network Design Nonconvex MINLP Discrete: Pipe connections/sizes Nonlinear: Pressure Loss 3 Pooling/Petrochemical Nonconvex MINLP Discrete: Which process to use? Nonlinear: Product Blending (among others) 24 / 126
40 Some MINLP Applications 1 Supply Chain Convex MINLP Discrete: Fixed charges for opening facilities Nonlinear: Nonlinear transportation costs 2 Gas/Water Network Design Nonconvex MINLP Discrete: Pipe connections/sizes Nonlinear: Pressure Loss 3 Pooling/Petrochemical Nonconvex MINLP Discrete: Which process to use? Nonlinear: Product Blending (among others) Luedtke Theorem 97.5% of all practical MINLPs have integer variables for one of two purposes... 1 Binary variables to make a multiple choice selection 2 Binary indicator variables that turn on/off continuous variables and/or constraints. 24 / 126
41 Supply Chain Outline 1 Introduction 2 Supply Chain Water Network Design Pooling 3 Algorithms Convex 4 Algorithms Nonconvex 5 The Future of MINLP? 25 / 126
42 Supply Chain Uncapacitated Facility Location Problem introduced by Günlük, Lee, and Weismantel ( 07) and classes of strong cutting planes derived 26 / 126
43 Supply Chain Uncapacitated Facility Location Problem introduced by Günlük, Lee, and Weismantel ( 07) and classes of strong cutting planes derived M: Facility N: Customer x ij : percentage of customer j N demand met by facility i M z i = 1 facility i M is built Fixed cost for opening facility i M Quadratic cost for meeting demand j N from facility i M 26 / 126
44 Supply Chain Quadratic Uncapacitated Facility Location Problem A very simple MIQP subject to z def = min c i z i + i M i M q ij xij 2 j N x ij z i i M, j N x ij = 1 j N i M x ij 0 i M, j N z i {0, 1} i M 27 / 126
45 Supply Chain Base Case of Induction to Luedtke Theorem Note that binary variables are used as indicators: z i = 0 x ij = 0 If z = 1, then we need to model the epigraph of x 2 ij 28 / 126
46 Supply Chain Base Case of Induction to Luedtke Theorem Note that binary variables are used as indicators: z i = 0 x ij = 0 If z = 1, then we need to model the epigraph of x 2 ij Intager Programmerz r Smrt Mixed Integer Linear Programmers carefully study simple problem structures to come up with good formulations for problems Good formulations closely approximate convex hull of feasible solutions Important to understand the structure of a special MINLP with indicator variables 28 / 126
47 Supply Chain A Very Simple Structure R def = { } (x, y, z) R 2 B y x 2, 0 x uz y z = 1 z x 29 / 126
48 Supply Chain A Very Simple Structure R def = { } (x, y, z) R 2 B y x 2, 0 x uz y z = 0 x = 0, y 0 z = 1 x u, y x 2 z z = 1 x 29 / 126
49 Supply Chain A Very Simple Structure R def = { } (x, y, z) R 2 B y x 2, 0 x uz y z = 0 x = 0, y 0 z = 1 x u, y x 2 z z = 1 Deep Insights conv(r) line connecting (0, 0, 0) to y = x 2 in the z = 1 plane x 29 / 126
50 Supply Chain Characterization of Convex Hull Deep Theorem #1 { } R = (x, y, z) R 2 B y x 2, 0 x uz conv(r) = { (x, y, z) R 3 yz x 2, 0 x uz, 0 z 1, y 0 } 30 / 126
51 Supply Chain Characterization of Convex Hull Deep Theorem #1 { } R = (x, y, z) R 2 B y x 2, 0 x uz conv(r) = { (x, y, z) R 3 yz x 2, 0 x uz, 0 z 1, y 0 } x 2 yz, y, z 0 Second Order Cone Programming There are effective, robust algorithms for optimizing over conv(r) 30 / 126
52 Supply Chain In General Giving You Some Perspective For a convex function f : R n R, the perspective function P : R n+1 R of f is { P(x, z) def 0 if z = 0 = zf (x/z) if z > 0 The epigraph of P(x, z) is a cone pointed at the origin whose lower shape is f (x) 31 / 126
53 Supply Chain In General Giving You Some Perspective For a convex function f : R n R, the perspective function P : R n+1 R of f is { P(x, z) def 0 if z = 0 = zf (x/z) if z > 0 The epigraph of P(x, z) is a cone pointed at the origin whose lower shape is f (x) Exploiting Your Perspective If z i is an indicator that the (nonlinear, convex) inequality f (x) 0 must hold, (otherwise x = 0), replace the inequality with its perspective version: z i f (x/z i ) 0 The resulting (convex) inequality is a much tighter relaxation of the feasible region. 31 / 126
54 Supply Chain Perspectivizing Quadratic Uncapacitated Facility Location z def = min c i z i + i M i M q ij xij 2 j N Steps x ij z i i M, j N x ij = 1 j N i M x ij 0 i M, j N z i {0, 1} i M 32 / 126
55 Supply Chain Perspectivizing Quadratic Uncapacitated Facility Location z def = min c i z i + i M i M x ij z i x ij = 1 i M x ij 0 q ij y ij j N i M, j N j N i M, j N z i {0, 1} i M x 2 ij y ij 0 i M, j N Steps 1 Make Objective Linear 32 / 126
56 Supply Chain Perspectivizing Quadratic Uncapacitated Facility Location z def = min c i z i + i M i M x ij z i x ij = 1 i M x ij 0 q ij y ij j N i M, j N j N i M, j N z i {0, 1} i M x 2 ij z i y ij 0 i M, j N Steps 1 Make Objective Linear 2 Apply Perspective 32 / 126
57 Supply Chain Perspectivizing Quadratic Uncapacitated Facility Location z def = min c i z i + i M i M x ij z i x ij = 1 i M x ij 0 q ij y ij j N i M, j N j N i M, j N z i {0, 1} i M x 2 ij z i y ij 0 i M, j N Steps 1 Make Objective Linear 2 Apply Perspective 3 Make Giant Profit 32 / 126
58 Supply Chain Strength of Relaxations z R : Value of NLP relaxation z GLW : Value of NLP relaxation after GLW cuts z P : Value of perspective relaxation z : Optimal solution value 33 / 126
59 Supply Chain Strength of Relaxations z R : Value of NLP relaxation z GLW : Value of NLP relaxation after GLW cuts z P : Value of perspective relaxation z : Optimal solution value M N z R z GLW z P z / 126
60 Supply Chain Strength of Relaxations z R : Value of NLP relaxation z GLW : Value of NLP relaxation after GLW cuts z P : Value of perspective relaxation z : Optimal solution value M N z R z GLW z P z / 126
61 Supply Chain Strength of Relaxations z R : Value of NLP relaxation z GLW : Value of NLP relaxation after GLW cuts z P : Value of perspective relaxation z : Optimal solution value M N z R z GLW z P z Cheers! 33 / 126
62 Supply Chain Impact of SOCP m = 30, n = 100 Convex MINLP, Original: CPU seconds, nodes Convex MINLP, w/glw Ineq.: CPU seconds, nodes 34 / 126
63 Supply Chain Impact of SOCP m = 30, n = 100 Convex MINLP, Original: CPU seconds, nodes Convex MINLP, w/glw Ineq.: CPU seconds, nodes MISOCP, Perspective, 23 CPU seconds, 44 B&B nodes 34 / 126
64 Supply Chain Impact of SOCP m = 30, n = 100 Convex MINLP, Original: CPU seconds, nodes Convex MINLP, w/glw Ineq.: CPU seconds, nodes MISOCP, Perspective, 23 CPU seconds, 44 B&B nodes Larger Instances m n T N Poifect 34 / 126
65 Water Network Design Outline 1 Introduction 2 Supply Chain Water Network Design Pooling 3 Algorithms Convex 4 Algorithms Nonconvex 5 The Future of MINLP? 35 / 126
66 Water Network Design Design of Water Distribution Networks Goal: Design minimum cost network from discrete pipe diameters to meet water demand 36 / 126
67 Water Network Design Design of Water Distribution Networks Goal: Design minimum cost network from discrete pipe diameters to meet water demand Sets N nodes in network S source nodes A: arcs in the network K: Pipe diameters 36 / 126
68 Water Network Design Design of Water Distribution Networks Variables q ij : flow pipe (i, j) A h i : hydraulic head at node i N d ij : diameter of pipe (i, j) A, where d ij {P 1,..., P r } a ij : area of cross section (i, j) A y ijk : Binary variable = 1 if pipe of size k K chosen for (i, j) A 37 / 126
69 Water Network Design Design of Water Distribution Networks Variables q ij : flow pipe (i, j) A h i : hydraulic head at node i N d ij : diameter of pipe (i, j) A, where d ij {P 1,..., P r } a ij : area of cross section (i, j) A y ijk : Binary variable = 1 if pipe of size k K chosen for (i, j) A Constraints Conservation of flow q ij (i,j) A (j,i) A q ji = D i, i N \ S. Flow bounds: (linear in a, but since a ij = 0.25πdij 2, nonlinear in d) V max a ij q ij V max a ij, (i, j) A. 37 / 126
70 Water Network Design Design of Water Distribution Networks Discrete pipe sizes d ij {P 1,..., P r } Use multiple choice integrality to remove nonlinearity a ij = πd 2 ij /4 This is in general a good modeling practice 38 / 126
71 Water Network Design Design of Water Distribution Networks Discrete pipe sizes d ij {P 1,..., P r } Use multiple choice integrality to remove nonlinearity a ij = πd 2 ij /4 This is in general a good modeling practice 1 Binary variables y ijk {0, 1} for k K 2 Model discrete choice as k K y ijk = 1, and k K 3 Model a ij = πdij 2 /4 (nonlinear) as a ij = k K ( π 4 P k P k y ijk = d ij. (i, j) A ) y ijk (i, j) A 38 / 126
72 Water Network Design Design of Water Distribution Networks Nonsmooth, nonlinear pressure loss as a function of flow q ij along arc (i, j) A and diameter d ij : h i h j = κ ij d c 1 ij sgn(q ij ) q ij c 2 (i, j) A 39 / 126
73 Water Network Design Design of Water Distribution Networks Nonsmooth, nonlinear pressure loss as a function of flow q ij along arc (i, j) A and diameter d ij : h i h j = κ ij d c 1 ij sgn(q ij ) q ij c 2 (i, j) A Another trick: disaggregate flow into flow variables for each pipe size: q ij = k K q ijk ( V max a ij )y ijk q ijk (V max a ij )y ijk 39 / 126
74 Water Network Design Why On Earth? Doing this allows us to write the (nonlinear) headloss as a function of a single variable q ijk Often in applications, the nonlinearity is low-dimensional (low {1, 2}) f (q) = sgn(q) q c 2 q 40 / 126
75 Water Network Design Why On Earth? Doing this allows us to write the (nonlinear) headloss as a function of a single variable q ijk Often in applications, the nonlinearity is low-dimensional (low {1, 2}) Therefore a piecewise linear approximation (or relaxation) may be sufficiently accurate f (q) = sgn(q) q c 2 q 40 / 126
76 Water Network Design Why On Earth? Doing this allows us to write the (nonlinear) headloss as a function of a single variable q ijk Often in applications, the nonlinearity is low-dimensional (low {1, 2}) Therefore a piecewise linear approximation (or relaxation) may be sufficiently accurate Luedtke s Theorem! Note also that y ijk = 0 q ijk = 0 f (q) = sgn(q) q c 2 q 40 / 126
77 Water Network Design Modeling Piecewise-Linear Functions Take Your Pick SOS2 Model [Beale and Tomlin, 1970, Beale and Forrest, 1976] Incremental Model [Markowitz and Manne, 1957] Multiple Choice Model [Jeroslow and Lowe, 1984] Convex Combination Model [Dantzig, 1960, Padberg, 2000] Disaggregated Convex Combination Model [Meyer, 1976] Logarithmic Model [Vielma and Nemhauser, 2011] 41 / 126
78 Water Network Design Multiple Choice Model Want to model ( i) f (q) = m i q + c i, q [B i 1, B i ] f (q) sgn(q) q c 2 B 0 B 1 B 2 B 3 B 4 q q = t = n w i, i=1 n (m i w i + c i b i ) i=1 B i 1 b i w i B i b i n 1 = i=1 b i i 1,..., n 42 / 126
79 Water Network Design Multiple Choice Model Want to model ( i) f (q) = m i q + c i, q [B i 1, B i ] f (q) sgn(q) q c 2 Introduce New Variables w i : = q if q [B i 1, B i ] b i : = 1 if q [B i 1, B i ] t f (q) B 0 B 1 B 2 B 3 B 4 q q = t = n w i, i=1 n (m i w i + c i b i ) i=1 B i 1 b i w i B i b i n 1 = i=1 b i i 1,..., n 42 / 126
80 Water Network Design Multiple Choice Model Want to model ( i) f (q) = m i q + c i, q [B i 1, B i ] f (q) sgn(q) q c 2 Introduce New Variables w i : = q if q [B i 1, B i ] b i : = 1 if q [B i 1, B i ] t f (q) B 0 B 1 B 2 B 3 B 4 q q = t = n w i, i=1 n (m i w i + c i b i ) i=1 B i 1 b i w i B i b i n 1 = i=1 b i i 1,..., n B 0 y q B n y 42 / 126
81 Water Network Design A Simple Trick Instead of modeling the piecewise-linear indicator in the standard way: B 0 y q B n y, 1 = n i=1 b i 43 / 126
82 Water Network Design A Simple Trick Instead of modeling the piecewise-linear indicator in the standard way: B 0 y q B n y, 1 = n i=1 b i Model it as y = n i=1 b i 43 / 126
83 Water Network Design A Simple Trick Instead of modeling the piecewise-linear indicator in the standard way: B 0 y q B n y, 1 = n i=1 b i Model it as y = n i=1 b i Resulting formulation is provably stronger (locally-ideal) All piecewise-linear functions turned on/off by an indicator have a similar modeling trick [Sridhar et al., 2013] It is the exact extension of the perspective reformulation to this nonconvex case 43 / 126
84 Pooling Outline 1 Introduction 2 Supply Chain Water Network Design Pooling 3 Algorithms Convex 4 Algorithms Nonconvex 5 The Future of MINLP? 44 / 126
85 Pooling The Pooling Problem The Pooling Problem A canonical problem in the petrochemical industry Decide how to blend (pool) feeds together to make products meeting quality requirements at minimum cost Feeds I Pools J Products K Generally each product has multiple attributes 45 / 126
86 Pooling Pooling Problem Basic variables x ij : Flow of input i to pool j x ik : Flow of input i to product k x jk : Flow from pool j to product k Key Parameters λ i : Concentration of attribute in feed i l k, υ k : Bounds on attribute concentration in product k C i, C j, C k : Capacities on feeds, pools, and outputs U ij, U ik, U kj : Flow capacities (some may be zero) Simple (Linear) Constraints Linear Constraints: Flow balance, bounds on flows, feed/pool/output capacity 46 / 126
87 Pooling Modeling the Quality Constraints Q-formulation [Ben-Tal et al., 1994] New variables q ij : proportion of flow to pool j coming from feed i: q ij = 1 i I q ij satisfy the relations: q ij = x ij t I x tj x ij = q ij t I x tj 47 / 126
88 Pooling Modeling the Quality Constraints Q-formulation [Ben-Tal et al., 1994] New variables q ij : proportion of flow to pool j coming from feed i: q ij = 1 q ij satisfy the relations: i I q ij = x ij t I x tj x ij = q ij t I x tj Reformulate using flow balance at pool j: t I x tj = k K x jk x ij = q ij k K x jk 47 / 126
89 Pooling Modeling the Quality Constraints Q-formulation [Ben-Tal et al., 1994] New variables q ij : proportion of flow to pool j coming from feed i: q ij = 1 q ij satisfy the relations: i I q ij = x ij t I x tj x ij = q ij t I x tj Reformulate using flow balance at pool j: t I x tj = k K x jk x ij = q ij k K Concentration of attribute at pool i is then: i I λ iq ij x jk 47 / 126
90 Pooling Modeling the Quality Constraints (2) Concentration of attribute at pool i: i I λ iq ij Quality constraint on product k: i I l k λ ix ik + j J i I λ iq ij x jk i I x ik + j J x jk υ k Multiply and gather terms (e.g., for upper bound): (λ i υ k )x ik + (λ i υ k )q ij x jk 0 i I j J i I 48 / 126
91 Pooling Modeling the Quality Constraints (2) Concentration of attribute at pool i: i I λ iq ij Quality constraint on product k: i I l k λ ix ik + j J i I λ iq ij x jk i I x ik + j J x jk υ k Multiply and gather terms (e.g., for upper bound): (λ i υ k )x ik + (λ i υ k )q ij x jk 0 i I j J i I Introduce variables w ijk to represent bilinear terms q ij x jk : x ij = q ij x jk = w ijk k K k K 48 / 126
92 Pooling Modeling the Quality Constraints (3) Bringing it all together q ij = 1, j J i I x ij = w ijk, i I, j J k K (λ i υ k )x ik + (λ i υ k )w ijk 0, k K i I j J i I (l k λ i )x ik + (l k λ i )w ijk 0 k K i I j J i I w ijk = q ij x jk, i I, j J, k K 49 / 126
93 Pooling PQ-Formulation [Tawarmalani and Sahinidis, 2002] 1 For each j J, multiply the constraint i I q ij = 1 with x jk : q ij x jk = x jk i I 2 For each j J, multiply the constraint: k K x jk C j with q ij : q ij x jk C j q ij k K 50 / 126
94 Pooling PQ-Formulation [Tawarmalani and Sahinidis, 2002] 1 For each j J, multiply the constraint i I q ij = 1 with x jk : q ij x jk = x jk i I 2 For each j J, multiply the constraint: k K x jk C j with q ij : q ij x jk C j q ij k K 3 Replace q ij x jk with w ijk : w ijk = x jk, i I w ijk C j q j k K 50 / 126
95 Pooling PQ-Formulation [Tawarmalani and Sahinidis, 2002] 1 For each j J, multiply the constraint i I q ij = 1 with x jk : q ij x jk = x jk i I 2 For each j J, multiply the constraint: k K x jk C j with q ij : q ij x jk C j q ij k K 3 Replace q ij x jk with w ijk : w ijk = x jk, i I w ijk C j q j k K Much better relaxations! Takeaway: Adding redundant constraints can significantly improve relaxations of nonconvex MINLPs. 50 / 126
96 Pooling The Pooling Problem with Binary Variables Arcs connecting the input streams to other nodes may be used only by paying a fixed cost. Network Design Ship assignment 51 / 126
97 Pooling The Pooling Problem with Binary Variables Arcs connecting the input streams to other nodes may be used only by paying a fixed cost. Network Design Ship assignment Add binary variables z ij for these arcs: x ij U ij z ij A real nonconvex, integer problem. 51 / 126
98 Pooling Practical Advice About the Luedtke Theorem We have seen many cases where we have continuous variable 0 x U and associated binary variable z {0, 1}. We see many models that write the expression w = zx. 52 / 126
99 Pooling Practical Advice About the Luedtke Theorem We have seen many cases where we have continuous variable 0 x U and associated binary variable z {0, 1}. We see many models that write the expression w = zx. On The Horror! Never multiply a binary variable by a bounded continuous variable You can linearize this: 0 w Uz U(1 z) x w U(1 z) 52 / 126
100 Pooling Practical Advice About the Luedtke Theorem (2) The same trick applies to binary variables that indicate constraints. Optimal transmission switching. Instead of x ij = α ij z ij (θ j θ i ) (i, j) E Write α ij (θ i θ j ) x ij + M(1 z ij ) 0 α ij (θ i θ j ) x ij M(1 z ij ) 0 (i, j) E (i, j) E Fancier models (disjunctive programming) exist [Grossmann and Lee, 2003]), but you have our permission to use big-m in this case 53 / 126
101 Pooling Other MINLP Applications 1 Portfolio Management Convex MIQP Discrete: Trading Strategy Nonlinear: Utility 2 Reactor Core Reloading Nonconvex MINLP Discrete: Bundle Reordering Nonlinear: Neutron Transport 3 Power Grid Operation Nonconvex MINLP Discrete: Unit Committment Nonlinear: AC Power Flow 4 Building Co-Generation Nonconvex MINLP Discrete: Unit Committment Nonlinear: Ramping Constraints 5 Oil-Spill Response Nonconvex MINLP Discrete: Resource Allocation Nonlinear: Fluid Flow & Chemistry 2. Core-Reloading 4. Building Demand 5. Oil-Spill Response 54 / 126
Mixed Integer Nonlinear Programming
Mixed Integer Nonlinear Programming IMA New Directions Short Course on Mathematical Optimization Jeff Linderoth and Jim Luedtke Department of Industrial and Systems Engineering University of Wisconsin-Madison
More informationSolving 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 informationMixed 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 informationSoftware 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 informationPERSPECTIVE REFORMULATION AND APPLICATIONS
PERSPECTIVE REFORMULATION AND APPLICATIONS OKTAY GÜNLÜK AND JEFF LINDEROTH Abstract. In this paper we survey recent work on the perspective reformulation approach that generates tight, tractable relaxations
More informationApplications and algorithms for mixed integer nonlinear programming
Applications and algorithms for mixed integer nonlinear programming Sven Leyffer 1, Jeff Linderoth 2, James Luedtke 2, Andrew Miller 3, Todd Munson 1 1 Mathematics and Computer Science Division, Argonne
More informationMINLP: 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 informationComputer Sciences Department
Computer Sciences Department Valid Inequalities for the Pooling Problem with Binary Variables Claudia D Ambrosio Jeff Linderoth James Luedtke Technical Report #682 November 200 Valid Inequalities for the
More informationA 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 informationSome 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 informationA 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 informationMixed-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 informationBasic 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 informationMILP reformulation of the multi-echelon stochastic inventory system with uncertain demands
MILP reformulation of the multi-echelon stochastic inventory system with uncertain demands Axel Nyberg Åbo Aademi University Ignacio E. Grossmann Dept. of Chemical Engineering, Carnegie Mellon University,
More informationFrom 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 informationOptimization in Process Systems Engineering
Optimization in Process Systems Engineering M.Sc. Jan Kronqvist Process Design & Systems Engineering Laboratory Faculty of Science and Engineering Åbo Akademi University Most optimization problems in production
More informationHeuristics 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 informationDevelopment 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 informationOn 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 informationIndicator 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 informationProjected Perspective Reformulations for NonLinear Network Design Problems
Projected Perspective Reformulations for NonLinear Network Design Problems Antonio Frangioni Claudio Gentile Enrico Grande Andrea Pacifici Dipartimento di Informatica, Università di Pisa Polo Universitario
More informationTruss Topology Optimization under Constraints. on Number of Different Design Variables
Truss Topology Optimization under Constraints on Number of Different Design Variables Yoshihiro Kanno (Tokyo Institute of Technology) (University of Tokyo) June 11, 2015 constraint on # of different design
More informationGLOBAL 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 informationOn Valid Inequalities for Quadratic Programming with Continuous Variables and Binary Indicators
On Valid Inequalities for Quadratic Programming with Continuous Variables and Binary Indicators Hongbo Dong and Jeff Linderoth Wisconsin Institutes for Discovery University of Wisconsin-Madison, USA, hdong6,linderoth}@wisc.edu
More informationValid Inequalities for Optimal Transmission Switching
Valid Inequalities for Optimal Transmission Switching Hyemin Jeon Jeff Linderoth Jim Luedtke Dept. of ISyE UW-Madison Burak Kocuk Santanu Dey Andy Sun Dept. of ISyE Georgia Tech 19th Combinatorial Optimization
More informationMinotaur: A Mixed-Integer Nonlinear Optimization Toolkit
ARGONNE NATIONAL LABORATORY 9700 South Cass Avenue Lemont, Illinois 60439 Minotaur: A Mixed-Integer Nonlinear Optimization Toolkit Ashutosh Mahajan, Sven Leyffer, Jeff Linderoth, James Luedtke, and Todd
More informationSolving 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 informationConvex 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 informationAdvances 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 informationThe 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 informationComputational Integer Programming. Lecture 2: Modeling and Formulation. Dr. Ted Ralphs
Computational Integer Programming Lecture 2: Modeling and Formulation Dr. Ted Ralphs Computational MILP Lecture 2 1 Reading for This Lecture N&W Sections I.1.1-I.1.6 Wolsey Chapter 1 CCZ Chapter 2 Computational
More informationA note on : A Superior Representation Method for Piecewise Linear Functions by Li, Lu, Huang and Hu
A note on : A Superior Representation Method for Piecewise Linear Functions by Li, Lu, Huang and Hu Juan Pablo Vielma, Shabbir Ahmed and George Nemhauser H. Milton Stewart School of Industrial and Systems
More informationSolving 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 informationMultiperiod 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 informationMixed Integer Nonlinear Programming
ROADEF 2014 Bordeaux 26 févier 2014. 1 c 2014 IBM corportation Decision Optimization Mixed Integer Nonlinear Programming algorithms Pierre Bonami (and many co-authors) IBM ILOG CPLEX "The mother of all
More informationImproved Big-M Reformulation for Generalized Disjunctive Programs
Improved Big-M Reformulation for Generalized Disjunctive Programs Francisco Trespalacios and Ignacio E. Grossmann Department of Chemical Engineering Carnegie Mellon University, Pittsburgh, PA 15213 Author
More informationImproved 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 informationOnline generation via offline selection - Low dimensional linear cuts from QP SDP relaxation -
Online generation via offline selection - Low dimensional linear cuts from QP SDP relaxation - Radu Baltean-Lugojan Ruth Misener Computational Optimisation Group Department of Computing Pierre Bonami Andrea
More informationThe Pooling Problem - Applications, Complexity and Solution Methods
The Pooling Problem - Applications, Complexity and Solution Methods Dag Haugland Dept. of Informatics University of Bergen Norway EECS Seminar on Optimization and Computing NTNU, September 30, 2014 Outline
More informationA 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, Turku,
More informationOn 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 informationConstraint Qualification Failure in Action
Constraint Qualification Failure in Action Hassan Hijazi a,, Leo Liberti b a The Australian National University, Data61-CSIRO, Canberra ACT 2601, Australia b CNRS, LIX, Ecole Polytechnique, 91128, Palaiseau,
More informationFeasibility 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 informationSemidefinite Programming Basics and Applications
Semidefinite Programming Basics and Applications Ray Pörn, principal lecturer Åbo Akademi University Novia University of Applied Sciences Content What is semidefinite programming (SDP)? How to represent
More informationMixed Integer Programming Models for Non-Separable Piecewise Linear Cost Functions
Mixed Integer Programming Models for Non-Separable Piecewise Linear Cost Functions Juan Pablo Vielma H. Milton Stewart School of Industrial and Systems Engineering Georgia Institute of Technology Joint
More informationA Lifted Linear Programming Branch-and-Bound Algorithm for Mixed Integer Conic Quadratic Programs
A Lifted Linear Programming Branch-and-Bound Algorithm for Mied Integer Conic Quadratic Programs Juan Pablo Vielma Shabbir Ahmed George L. Nemhauser H. Milton Stewart School of Industrial and Systems Engineering
More informationMIXED INTEGER PROGRAMMING APPROACHES FOR NONLINEAR AND STOCHASTIC PROGRAMMING
MIXED INTEGER PROGRAMMING APPROACHES FOR NONLINEAR AND STOCHASTIC PROGRAMMING A Thesis Presented to The Academic Faculty by Juan Pablo Vielma Centeno In Partial Fulfillment of the Requirements for the
More informationA note on : A Superior Representation Method for Piecewise Linear Functions
A note on : A Superior Representation Method for Piecewise Linear Functions Juan Pablo Vielma Business Analytics and Mathematical Sciences Department, IBM T. J. Watson Research Center, Yorktown Heights,
More informationMixed-Integer Nonlinear Decomposition Toolbox for Pyomo (MindtPy)
Mario R. Eden, Marianthi Ierapetritou and Gavin P. Towler (Editors) Proceedings of the 13 th International Symposium on Process Systems Engineering PSE 2018 July 1-5, 2018, San Diego, California, USA 2018
More informationRelaxations of multilinear convex envelopes: dual is better than primal
of multilinear convex envelopes: dual is better than primal 1 LIX, École Polytechnique, Palaiseau, France June 7th, 2012 11 th International Symposium on Experimental Algorithms (SEA) 2012 - Bordeaux (France)
More informationExact Computation of Global Minima of a Nonconvex Portfolio Optimization Problem
Frontiers In Global Optimization, pp. 1-2 C. A. Floudas and P. M. Pardalos, Editors c 2003 Kluwer Academic Publishers Exact Computation of Global Minima of a Nonconvex Portfolio Optimization Problem Josef
More informationSolving LP and MIP Models with Piecewise Linear Objective Functions
Solving LP and MIP Models with Piecewise Linear Obective Functions Zonghao Gu Gurobi Optimization Inc. Columbus, July 23, 2014 Overview } Introduction } Piecewise linear (PWL) function Convex and convex
More information19. Fixed costs and variable bounds
CS/ECE/ISyE 524 Introduction to Optimization Spring 2017 18 19. Fixed costs and variable bounds ˆ Fixed cost example ˆ Logic and the Big M Method ˆ Example: facility location ˆ Variable lower bounds Laurent
More informationBCOL 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 informationMixed Integer Nonlinear Programming (MINLP)
Tutorial: Mixed Integer Nonlinear Programming () Sven Leyffer MCS Division Argonne National Lab leyffer@mcs.anl.gov Jeff Linderoth ISE Department Lehigh University jtl3@lehigh.edu informs Annual Meeting
More informationGestion de la production. Book: Linear Programming, Vasek Chvatal, McGill University, W.H. Freeman and Company, New York, USA
Gestion de la production Book: Linear Programming, Vasek Chvatal, McGill University, W.H. Freeman and Company, New York, USA 1 Contents 1 Integer Linear Programming 3 1.1 Definitions and notations......................................
More informationConvexification of Mixed-Integer Quadratically Constrained Quadratic Programs
Convexification of Mixed-Integer Quadratically Constrained Quadratic Programs Laura Galli 1 Adam N. Letchford 2 Lancaster, April 2011 1 DEIS, University of Bologna, Italy 2 Department of Management Science,
More informationMachine 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 informationAn exact reformulation algorithm for large nonconvex NLPs involving bilinear terms
An exact reformulation algorithm for large nonconvex NLPs involving bilinear terms Leo Liberti CNRS LIX, École Polytechnique, F-91128 Palaiseau, France. E-mail: liberti@lix.polytechnique.fr. Constantinos
More informationOptimization of a Nonlinear Workload Balancing Problem
Optimization of a Nonlinear Workload Balancing Problem Stefan Emet Department of Mathematics and Statistics University of Turku Finland Outline of the talk Introduction Some notes on Mathematical Programming
More informationDisconnecting 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 informationCONSTRAINED OPTIMIZATION LARS IMSLAND
Introduction to course TTK16 Mixed integer optimization in energy and oil and gas systems CONSTRAINED OPTIMIZATION LARS IMSLAND TTK16 overview This introduction lecture One week intensive course Starting
More informationExtended Formulations, Lagrangian Relaxation, & Column Generation: tackling large scale applications
Extended Formulations, Lagrangian Relaxation, & Column Generation: tackling large scale applications François Vanderbeck University of Bordeaux INRIA Bordeaux-Sud-Ouest part : Defining Extended Formulations
More informationValid Inequalities for Potential-Constrained Network Design
Valid Inequalities for Potential-Constrained Network Design Hyemin Jeon Jeff Linderoth Jim Luedtke Dept. of ISyE UW-Madison jrluedt1@wisc.edu MINLP 2014, CMU, Pittsburgh, PA June 3, 2014 Jeon, Linderoth,
More informationAPPLIED MECHANISM DESIGN FOR SOCIAL GOOD
APPLIED MECHANISM DESIGN FOR SOCIAL GOOD JOHN P DICKERSON Lecture #4 09/08/2016 CMSC828M Tuesdays & Thursdays 12:30pm 1:45pm PRESENTATION LIST IS ONLINE! SCRIBE LIST COMING SOON 2 THIS CLASS: (COMBINATORIAL)
More informationMixed 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 informationLift-and-project cuts for convex mixed integer nonlinear programs. Linear programming based separation and extended formulations
Math. Prog. Comp. DOI 10.1007/s12532-017-0118-1 FULL LENGTH PAPER 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Lift-and-project cuts for convex mixed integer nonlinear programs Linear programming based separation
More informationSemidefinite Relaxations for Non-Convex Quadratic Mixed-Integer Programming
Semidefinite Relaxations for Non-Convex Quadratic Mixed-Integer Programming Christoph Buchheim 1 and Angelika Wiegele 2 1 Fakultät für Mathematik, Technische Universität Dortmund christoph.buchheim@tu-dortmund.de
More informationNonconvex Quadratic Programming: Return of the Boolean Quadric Polytope
Nonconvex Quadratic Programming: Return of the Boolean Quadric Polytope Kurt M. Anstreicher Dept. of Management Sciences University of Iowa Seminar, Chinese University of Hong Kong, October 2009 We consider
More informationUsing 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 informationCutting Planes for First Level RLT Relaxations of Mixed 0-1 Programs
Cutting Planes for First Level RLT Relaxations of Mixed 0-1 Programs 1 Cambridge, July 2013 1 Joint work with Franklin Djeumou Fomeni and Adam N. Letchford Outline 1. Introduction 2. Literature Review
More information18 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 informationConvex Optimization. (EE227A: UC Berkeley) Lecture 28. Suvrit Sra. (Algebra + Optimization) 02 May, 2013
Convex Optimization (EE227A: UC Berkeley) Lecture 28 (Algebra + Optimization) 02 May, 2013 Suvrit Sra Admin Poster presentation on 10th May mandatory HW, Midterm, Quiz to be reweighted Project final report
More informationEasy and not so easy multifacility location problems... (In 20 minutes.)
Easy and not so easy multifacility location problems... (In 20 minutes.) MINLP 2014 Pittsburgh, June 2014 Justo Puerto Institute of Mathematics (IMUS) Universidad de Sevilla Outline 1 Introduction (In
More informationMultiobjective 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 informationAnalyzing 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 informationOn Handling Indicator Constraints in Mixed Integer Programming
On Handling Indicator Constraints in Mixed Integer Programming Pietro Belotti 1, Pierre Bonami 2, Matteo Fischetti 3, Andrea Lodi 4, Michele Monaci 3, Amaya Nogales-Gómez 5, and Domenico Salvagnin 3 1
More informationDevelopment 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 informationA Principled Approach to Mixed Integer/Linear Problem Formulation
A Principled Approach to Mixed Integer/Linear Problem Formulation J N Hooker September 9, 2008 Abstract We view mixed integer/linear problem formulation as a process of identifying disjunctive and knapsack
More information15. Conic optimization
L. Vandenberghe EE236C (Spring 216) 15. Conic optimization conic linear program examples modeling duality 15-1 Generalized (conic) inequalities Conic inequality: a constraint x K where K is a convex cone
More informationOperations Research Methods in Constraint Programming
Operations Research Methods in Constraint Programming John Hooker Carnegie Mellon University Prepared for Lloret de Mar, Spain June 2007 2007 Slide 1 CP and OR Have Complementary Strengths CP: Inference
More informationA 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 informationModeling Disjunctive Constraints with a Logarithmic Number of Binary Variables and Constraints
Modeling Disjunctive Constraints with a Logarithmic Number of Binary Variables and Constraints Juan Pablo Vielma and George L. Nemhauser H. Milton Stewart School of Industrial and Systems Engineering,
More informationSecond-Order Cone Program (SOCP) Detection and Transformation Algorithms for Optimization Software
and Second-Order Cone Program () and Algorithms for Optimization Software Jared Erickson JaredErickson2012@u.northwestern.edu Robert 4er@northwestern.edu Northwestern University INFORMS Annual Meeting,
More informationOutline. 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 informationNumerical 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 informationStrong-Branching Inequalities for Convex Mixed Integer Nonlinear Programs
Computational Optimization and Applications manuscript No. (will be inserted by the editor) Strong-Branching Inequalities for Convex Mixed Integer Nonlinear Programs Mustafa Kılınç Jeff Linderoth James
More informationPseudo basic steps: Bound improvement guarantees from Lagrangian decomposition in convex disjunctive programming
Pseudo basic steps: Bound improvement guarantees from Lagrangian decomposition in convex disjunctive programming Dimitri J. Papageorgiou and Francisco Trespalacios Corporate Strategic Research ExxonMobil
More informationNetwork Flows. 6. Lagrangian Relaxation. Programming. Fall 2010 Instructor: Dr. Masoud Yaghini
In the name of God Network Flows 6. Lagrangian Relaxation 6.3 Lagrangian Relaxation and Integer Programming Fall 2010 Instructor: Dr. Masoud Yaghini Integer Programming Outline Branch-and-Bound Technique
More information23. 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 informationHeuristics 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 informationGlobal Optimization of Non-convex Generalized Disjunctive Programs: A Review on Relaxations and Solution Methods
Noname manuscript No. (will be inserted by the editor) Global Optimization of Non-convex Generalized Disjunctive Programs: A Review on Relaxations and Solution Methods Juan P. Ruiz Ignacio E. Grossmann
More informationIBM 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 informationOPTIMIZATION. joint course with. Ottimizzazione Discreta and Complementi di R.O. Edoardo Amaldi. DEIB Politecnico di Milano
OPTIMIZATION joint course with Ottimizzazione Discreta and Complementi di R.O. Edoardo Amaldi DEIB Politecnico di Milano edoardo.amaldi@polimi.it Website: http://home.deib.polimi.it/amaldi/opt-15-16.shtml
More informationLifting for conic mixed-integer programming
Math. Program., Ser. A DOI 1.17/s117-9-282-9 FULL LENGTH PAPER Lifting for conic mixed-integer programming Alper Atamtürk Vishnu Narayanan Received: 13 March 28 / Accepted: 28 January 29 The Author(s)
More information3.7 Cutting plane methods
3.7 Cutting plane methods Generic ILP problem min{ c t x : x X = {x Z n + : Ax b} } with m n matrix A and n 1 vector b of rationals. According to Meyer s theorem: There exists an ideal formulation: conv(x
More informationConvex Optimization M2
Convex Optimization M2 Lecture 8 A. d Aspremont. Convex Optimization M2. 1/57 Applications A. d Aspremont. Convex Optimization M2. 2/57 Outline Geometrical problems Approximation problems Combinatorial
More informationOn Perspective Functions, Vanishing Constraints, and Complementarity Programming
On Perspective Functions, Vanishing Constraints, and Complementarity Programming Fast Mixed-Integer Nonlinear Feedback Control Christian Kirches 1, Sebastian Sager 2 1 Interdisciplinary Center for Scientific
More informationCan Li a, Ignacio E. Grossmann a,
A generalized Benders decomposition-based branch and cut algorithm for two-stage stochastic programs with nonconvex constraints and mixed-binary first and second stage variables Can Li a, Ignacio E. Grossmann
More informationCan Li a, Ignacio E. Grossmann a,
A generalized Benders decomposition-based branch and cut algorithm for two-stage stochastic programs with nonconvex constraints and mixed-binary first and second stage variables Can Li a, Ignacio E. Grossmann
More information