Solving mixed-integer partial differential equation constrained - - PowerPoint PPT Presentation

▶
solving mixed integer partial differential equation
SMART_READER_LITE
LIVE PREVIEW

Solving mixed-integer partial differential equation constrained - - PowerPoint PPT Presentation

Jun 22, 2023 23 likes •235 views

Solving mixed-integer partial differential equation constrained optimization Sven Leyffer 1 , Sebastian Sager 2 , and Anna Th unen 1 , 2 1 Argonne National Laboratory and 2 University of Magdeburg September 1, 2016 1 / 21 Problem Definition

slide-1
SLIDE 1

Solving mixed-integer partial differential equation constrained optimization

Sven Leyffer1, Sebastian Sager2, and Anna Th¨ unen1,2

1Argonne National Laboratory and 2University of Magdeburg

September 1, 2016

1 / 21

slide-2
SLIDE 2

Problem Definition and Motivation Approach: First Discretize then Optimize Every Day Problem . . . . . . Control of Heat Equation [Iftime and Demetriou(2009)] Solving MIPDECOs Summary Outlook

2 / 21

slide-3
SLIDE 3

Problem Definition and Motivation

Motivation: Physics combined with discrete choices, e.g.

◮ Topology optimization ◮ Energy (fossil & renewable) ◮ new problem class ⇒ new approaches

min

u,v,w F(u, v, w)

(integrals) s.t. C(u, v, w) = 0 (PDE+cond.) u ∈ U, v ∈ V and w ∈ W (function spaces), e.g. time and space dependency

◮ u(t, x, y, z): PDE states ◮ v(t, x, y, z): continuous controls ◮ w(t, x, y, z): binary (controls) are

infinite-dimensional

3 / 21

slide-4
SLIDE 4

Approach: First Discretize then Optimize

MIPDECO Discretize model Objective integrals, e.g. trapezoidal rule Constraints PDE, e.g. finite differences MINLP binaries may depend on discretization high dimensional Solve! State elimination Straight forward Rounding

  • ptimality & fast
  • ptimality & slow

good & faster

4 / 21

slide-5
SLIDE 5

Every Day Problem . . .

5 / 21

slide-6
SLIDE 6

. . . Control of Heat Equation [Iftime and Demetriou(2009)]

◮ adapted by Hante ISMP2015 ◮ Control temperature with actuator (= hairdryer) ◮ Select sequence of control inputs (actuators) over time ◮ Continuous control: intensity . . . heat/cool ◮ Match prescribed temperature profile (no fog) ◮ . . . de-fog bathroom mirror with hairdryer

6 / 21

slide-7
SLIDE 7

Control of Heat Equation: Mathematical Formulation

min

u,v,w

||u(tf )||2

2,Ω + 2||u||2 2,Ω×T +

1 500

9

  • l=1

||vl||2

2,T

  • bjective

s.t. ∂u ∂t − κ∆u =

9

  • l=1

vlfl in Ω × T heat eq u = 0

  • n ∂Ω × T

Dirichlet u(0) = 100 sin(πx) sin(πy) in Ω initial state − Mwl ≤ vl ≤ Mwl for all l ∈ {1, . . . , 9} in T big M

9

  • l=1

wl = 1 in T source budget wl(t) ∈ {0, 1} for all l ∈ {1, . . . , 9} in T binary.

◮ u(t, x, y) states (=temperature) ◮ vl(t) continous controls (=intensity) ◮ wl(t) binary controls (=location)

7 / 21

slide-8
SLIDE 8

How does the solution look like?

(Loading HeatAct32Rnd.avi)

8 / 21

slide-9
SLIDE 9

Solving MIPDECOs: Straight Forward MINLP

◮ discretize heat equation & other

constraints ⇒ MIQP

◮ build a model in AMPL ◮ apply preferred MIQP solver

(CPLEX, GUROBI, BONMIN, . . . ) ..easy peasy? Mesh size 8 16 32 64 Variables total 984 7728 63072 513216 binary 72 144 288 576 CPU (CPLEX) 7.3 5567.3 − −

9 / 21

slide-10
SLIDE 10

Solving MIPDECOs: Elimination of PDE

∂u ∂t − κ∆u =

9

  • l=1

vlfl u|∂Ω = 0 u(0) = 100 sin(πx) sin(πy) ⇒ AU = BV + d

◮ A: 5-point-stencil & implicit euler for heat equation ◮ U = vec(u): states, dim(U) = 2N2Tn ◮ B: gaussian source term f ◮ V = vec(v): continous controls, dim(V ) = 9Tn ◮ d: initial/boundary conditions ◮ we add binaries w later to assure that only in one location a

hairdryer is operated

10 / 21

slide-11
SLIDE 11

Solving MIPDECOs: Elimination of PDE (continued)

AU = BV + d ⇔ AU = B

9Tn

  • i=1

Viei + d ⇔ U =

9Tn

  • i=1

Vi A−1Bei

inhomogeneous part

+ A−1d

homogeneous part

  • 1. solve (9Tn + 1) linear systems

⇒ vec(¯ uhom.) = A−1d and vec(¯ ut,l) = A−1Bei

  • 2. solve reduced MIQP: ⇒ v∗

◮ no states, 9Tn continous (vt,l), 9Tn binary (wt,l) ◮ ¯

u data in objective

  • 3. construct the solution: u = v∗

t,l ¯

ut,l + ¯ uhom.

11 / 21

slide-12
SLIDE 12

Solving MIPDECOs: Elimination of PDE (continued)

Mesh size 8 16 32 Straight forward CPLEX 7.3 5567.3 − Elimination total 0.4 6.3 137.5 Linear systems (IPOPT) 0.1 3.3 123.2 MIQP (CPLEX) 0.3 2.9 14.3

  • Elim. with convolution total

1.0 3.4 16.8 Linear systems (IPOPT) 0.0 0.2 3.6 MIQP (CPLEX) 1.0 3.2 13.2

◮ optimality ◮ using convolution:

  • nly 9 instead of 9Tn linear

systems prior to optimization

◮ doesn’t work for nonlinear PDEs,

e.g. Helmholtz Eq

12 / 21

slide-13
SLIDE 13

Solving MIPDECOs: Sum-Up Rounding

◮ space discretized PDE

⇒ set of ODEs

◮ apply Sum-up Rounding

[Sager(2005)] for Optimal Control in every location Integer relaxation of MIQP ⇒ fractional ( ˜ w1, . . . , ˜ wTn) for every location For t = 1, . . . , Tn

◮ Compute rounding residual:

rt = ˜ wt +

t−1

  • τ=0

( ˜ wτ − wτ)

◮ Round:

wt = 1 if rt > 1

2

else End t 1 2 3 4 5 1 0.5 ˜ wt wt wt

13 / 21

slide-14
SLIDE 14

Solving MIPDECOs: Sum-Up Rounding (continued)

◮ do ODE-Sum-up Rounding in every location ◮ Theory for ODE Sum-up Rounding ◮ between proceeding to the next location we may resolve

(integer relaxation of remaining sub-MIQP)

◮ doesn’t satisfy the source budget (only one hairdryer) for all

t = 1, . . . , Tn:

9

  • l=1

wt,l = 1

◮ how to choose processing order of the locations? ◮ cheap, but large optimality gap

Mesh size #QPs 8 16 32 CPU time with resolve 9+1 0.03 1.74 55.86 CPU time without resolve 2 0.00 0.43 15.18

  • Opt. gap with resolve

9+1 0.66 0.75 0.66

  • Opt. gap without resolve

2 1.03 1.37 1.23

14 / 21

slide-15
SLIDE 15

Solving MIPDECOs: Knapsack-SUR Rounding

◮ SUR cheap, but large optimality gap ◮ Knapsack expensive, but small/no optimality gap

⇒ Combine! For t = 1, . . . , Tn

◮ %choose location largest continous control ¯

l = argmaxl|vt,l| (Knapsack Rounding)

◮ calculate rounding residual rt,l ◮ choose location largest residual ¯

l = argmaxl|rt,l|

◮ fix wt,¯ l = 1, all other wt,l = 0 ◮ solve integer relaxation of remaining sub-MIQP (resolve,

  • ptional)

End

15 / 21

slide-16
SLIDE 16

Solving MIPDECOs: Knapsack-SUR Rounding (continued)

◮ combines advantages of Knapsack and SUR Rounding ◮ preserves optimality of Knapsack in the resolved version ◮ reduce optimality gap in simple version

Mesh size (Tn) #QPs 8 16 32 CPU with resolve Tn+1 0.03 1.93 190.95 CPU without resolve 2 0.01 0.23 17.27

  • Opt. gap with resolve

Tn+1

  • Opt. gap without resolve

2 < 0.01 < 0.001 < 0.01

16 / 21

slide-17
SLIDE 17

Summary

Mesh size 8 16 32 #QPs CPU gap CPU gap CPU gap MIQP

  • 7.3
  • > 104
  • PDE

simple 9Tn LS 0.4

  • 6.3
  • 137.5
  • Elim.

convol. 9 LS 1.0

  • 3.4
  • 16.8
  • SUR

resolve 9+1 0.0 0.7 1.7 0.8 55.9 0.7 simple 2 0.0 1.0 0.4 1.4 15.2 1.2 Knap- resolve Tn+1 0.0

  • 2.5
  • 194.7
  • sack

simple 2 0.0 0.01 0.3 0.02 15.9 0.26 Knap. resolve Tn+1 0.0

  • 1.9
  • 191.0
  • SUR

simple 2 0.0 < 0.01 0.2 < 0.001 17.3 < 0.01

◮ PDE Elimination works well ◮ Knapsack-SUR as well ◮ Rounding possible even if PDE nonlinear

17 / 21

slide-18
SLIDE 18

Outlook

◮ develop theory ◮ move AMPL code to more efficient frameworks ◮ parallelization ◮ more challenging problems, such as nonlinear PDEs

Cloaking [Haslinger and M¨ akinen(2015)] (Helmholtz equation) (a) Scatterer (b) u without resolve (c) u with resolve

18 / 21

slide-19
SLIDE 19

References I

  • P. Belotti, C. Kirches, S. Leyffer, J. Linderoth, J. Luedtke, and
  • A. Mahajan.

Mixed-integer nonlinear optimization. Acta Numerica, 22:1–131, 5 2013. ISSN 1474-0508.

  • J. Haslinger and R. A. M¨

akinen. On a topology optimization problem governed by two-dimensional helmholtz equation.

  • Comput. Optim. Appl., 62(2):517–544, Nov. 2015.

ISSN 0926-6003.

  • O. V. Iftime and M. A. Demetriou.

Optimal control of switched distributed parameter systems with spatially scheduled actuators. Automatica, 45(2):312 – 323, 2009. ISSN 0005-1098.

19 / 21

slide-20
SLIDE 20

References II

  • S. Sager.

Numerical methods for mixed-integer optimal control problems. Der andere Verlag, T¨

  • nning, L¨

ubeck, Marburg, 2005. ISBN 3-89959-416-9.

  • F. Tr¨
  • ltzsch.

Optimal Control of Partial Differential Equations, volume 112

  • f Graduate Studies in Mathematics.

American Mathematical Society, Providence, 2010.

  • A. W¨

achter and T. L. Biegler. On the implementation of an interior-point filter line-search algorithm for large-scale nonlinear programming. Mathematical Programming, 106(1):25–57, 2006. ISSN 1436-4646.

20 / 21

slide-21
SLIDE 21

Pictures

https://s-media-cache- ak0.pinimg.com/236x/de/9d/17/de9d1746ab1a644a56f4e81d058a0f44.jpg http://images.travelpod.com/users/lduckett/2.1235847600.anti- fog-mirrorx.jpg https://travelqueentips.files.wordpress.com/2010/02/p2220006.jpg http://sharein.org/wp-content/uploads/2015/10/foggy-mirror- rectify-it-with-petrolium-jelly.jpg https://s-media-cache- ak0.pinimg.com/736x/0c/d3/f5/0cd3f502ce2afaaa2af0e6ab291d4c5b.jpg http://3.bp.blogspot.com/-TcTMQ63GeKQ/UZJYHW- nJGI/AAAAAAAAABE/5Rtke7XS1ug/s1600/PIKACHU-pikachu- 29274386-861-927.jpg http://orig05.deviantart.net/4594/f/2012/200/d/8/pikachu is - happy by otakuron-d57tnmg.png

21 / 21