MILP. LP: max cx ' MILP: some integer. ILP: x integer BLP: x 0,1. x 1. x 2 2 2, c ,

Similar documents
LARRY SNYDER DEPT. OF INDUSTRIAL AND SYSTEMS ENGINEERING CENTER FOR VALUE CHAIN RESEARCH LEHIGH UNIVERSITY

Kent State University CS 4/ Design and Analysis of Algorithms. Dept. of Math & Computer Science LECT-16. Dynamic Programming

INTRODUCTION INTRODUCTION. Moisès Graells Semi-continuous processes

LECTURE NOTES Duality Theory, Sensitivity Analysis, and Parametric Programming

Solving two-person zero-sum game by Matlab

TPL-Aware Displacement-driven Detailed Placement Refinement with Coloring Constraints

Modeling and Solving Nontraditional Optimization Problems Session 2a: Conic Constraints

TWO STAGE FACILITY LOCATION PROBLEM: LAGRANGIAN BASED HEURISTICS

CS 534: Computer Vision Model Fitting

Introduction to linear programming

Needed Information to do Allocation

5.3 Cutting plane methods and Gomory fractional cuts

GSLM Operations Research II Fall 13/14

3 INTEGER LINEAR PROGRAMMING

MVE165/MMG631 Linear and integer optimization with applications Lecture 9 Discrete optimization: theory and algorithms

Heuristics in Commercial MIP Solvers Part I (Heuristics in IBM CPLEX)

LP Rounding for k-centers with Non-uniform Hard Capacities

An Iterative Solution Approach to Process Plant Layout using Mixed Integer Optimisation

XLVII SIMPÓSIO BRASILEIRO DE PESQUISA OPERACIONAL

The SAS/OR s OPTMODEL Procedure :

MVE165/MMG630, Applied Optimization Lecture 8 Integer linear programming algorithms. Ann-Brith Strömberg

Fundamentals of Integer Programming

Smoothing Spline ANOVA for variable screening

Optimization Methods: Integer Programming Integer Linear Programming 1. Module 7 Lecture Notes 1. Integer Linear Programming

People have been thinking about network problems for a long time Koenigsberg Bridge problem (Euler, 1736)

A Facet Generation Procedure. for solving 0/1 integer programs

2 is not feasible if rounded. x =0,x 2

A mathematical programming approach to the analysis, design and scheduling of offshore oilfields

U.C. Berkeley CS294: Beyond Worst-Case Analysis Handout 5 Luca Trevisan September 7, 2017

MOBILE Cloud Computing (MCC) extends the capabilities

How Accurately Can We Model Timing In A Placement Engine?

11. APPROXIMATION ALGORITHMS

The SYMPHONY Callable Library for Mixed-Integer Linear Programming

Introduction to Mathematical Programming IE406. Lecture 20. Dr. Ted Ralphs

Discriminative Dictionary Learning with Pairwise Constraints

CS246: Mining Massive Datasets Jure Leskovec, Stanford University

Heuristics in MILP. Group 1 D. Assouline, N. Molyneaux, B. Morén. Supervisors: Michel Bierlaire, Andrea Lodi. Zinal 2017 Winter School

Outline. Modeling. Outline DMP204 SCHEDULING, TIMETABLING AND ROUTING. 1. Models Lecture 5 Mixed Integer Programming Models and Exercises

I. INTRODUCTION INFRASTRUCTURE providers (InPs), such as data center

An Optimal Algorithm for Prufer Codes *

An efficient iterative source routing algorithm

OPL: a modelling language

Greedy Technique - Definition

A HEURISTIC METHOD FOR RELIABILITY REDUNDANCY OPTIMIZATION OF FLOW NETWORKS

Concurrent Apriori Data Mining Algorithms

Noncommercial Software for Mixed-Integer Linear Programming

5 The Primal-Dual Method

Exploiting Degeneracy in MIP

On Mixed-Integer (Linear) Programming and its connection with Data Science

1 Introducton Effcent and speedy recovery of electrc power networks followng a major outage, caused by a dsaster such as extreme weather or equpment f

Intra-Parametric Analysis of a Fuzzy MOLP

The SYMPHONY Callable Library for Mixed-Integer Linear Programming

A NEW SEQUENTIAL CUTTING PLANE ALGORITHM FOR SOLVING MIXED INTEGER NONLINEAR PROGRAMMING PROBLEMS

Assignment # 2. Farrukh Jabeen Algorithms 510 Assignment #2 Due Date: June 15, 2009.

Biostatistics 615/815

2. Modeling AEA 2018/2019. Based on Algorithm Engineering: Bridging the Gap Between Algorithm Theory and Practice - ch. 2

Polyhedral Compilation Foundations

(Duality), Warm Starting, and Sensitivity Analysis for MILP

Parallelism for Nested Loops with Non-uniform and Flow Dependences

Compiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz

Sum of Linear and Fractional Multiobjective Programming Problem under Fuzzy Rules Constraints

Applied Mixed Integer Programming: Beyond 'The Optimum'

The MIP-Solving-Framework SCIP

A Lifted Linear Programming Branch-and-Bound Algorithm for Mixed Integer Conic Quadratic Programs

Quality Improvement Algorithm for Tetrahedral Mesh Based on Optimal Delaunay Triangulation

How to use your favorite MIP Solver: modeling, solving, cannibalizing. Andrea Lodi University of Bologna, Italy

Cost-efficient deployment of distributed software services

Exercises (Part 4) Introduction to R UCLA/CCPR. John Fox, February 2005

6.854 Advanced Algorithms Petar Maymounkov Problem Set 11 (November 23, 2005) With: Benjamin Rossman, Oren Weimann, and Pouya Kheradpour

Using COIN-OR to Solve the Uncapacitated Facility Location Problem

Efficient Video Coding with R-D Constrained Quadtree Segmentation

Routing on Switch Matrix Multi-FPGA Systems

Scheduling with Integer Time Budgeting for Low-Power Optimization

Efficient Load-Balanced IP Routing Scheme Based on Shortest Paths in Hose Model. Eiji Oki May 28, 2009 The University of Electro-Communications

Linking GAMS to Solvers Using COIN-OSI. Michael Bussieck Steve Dirkse GAMS Development Corporation

Design of Structure Optimization with APDL

MVE165/MMG631 Linear and integer optimization with applications Lecture 7 Discrete optimization models and applications; complexity

Maximum Variance Combined with Adaptive Genetic Algorithm for Infrared Image Segmentation

CMPS 10 Introduction to Computer Science Lecture Notes

Linear & Integer Programming: A Decade of Computation

Support Vector Machines

AADL : about scheduling analysis

The Codesign Challenge

The Gurobi Solver V1.0

DESIGN OF PLASTICS SEPARATION SYSTEMS UNDER UNCERTAINTY BY THE SAMPLE AVERAGE APPROXIMATION METHOD

The University of Jordan Department of Mathematics. Branch and Cut

Virtual Memory. Background. No. 10. Virtual Memory: concept. Logical Memory Space (review) Demand Paging(1) Virtual Memory

UC Berkeley Working Papers

Motivation for Heuristics

Programming in Fortran 90 : 2017/2018

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization

Vectorization in the Polyhedral Model

Meta-heuristics for Multidimensional Knapsack Problems

NGPM -- A NSGA-II Program in Matlab

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS

Optimal Scheduling of Capture Times in a Multiple Capture Imaging System

lpsymphony - Integer Linear Programming in R

Face Recognition University at Buffalo CSE666 Lecture Slides Resources:

Introduction to Mathematical Programming IE406. Lecture 9. Dr. Ted Ralphs

The Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique

Transcription:

MILP LP: max cx ' s.t. Ax b x 0 MILP: some nteger x max 6x 8x s.t. x x x 7 x, x 0 c A 6 8, 0 b 7 ILP: x nteger BLP: x 0, x 4 x, cx * * 0 4 5 6 x 06

Branch and Bound x 4 0 max 6x 8x s.t. xx x 7 x, x 0 x, x nteger 6 0 4 5 0 8 6 0 4 5 6 x c 6 8 A, 0 b 7 UB, LB 0 5 0 8 4 LP 6 UB, LB 0 x x 4 x x Incumbent UB, LB 6 UB, LB 6 x x UB 0, LB 6 Incumbent x x UB 0, LB 8 7 Incumbent UB 0, LB 0 gap 0 0 STOP Fathomed, nfeasble Fathomed, nfeasble 07

MILP Solvers 08

MILP Solvers x x Gomory cut Cplex (IBM, frst solver) Gurob (dev Robert Bxby) Xpress (used by LLamasoft) SAS/OR (part of SAS sys) Symphony (open source) Presolve: elmnate varables x x, x, x 0 and nteger x x 0 Cuttng planes: keeps all nteger solutons and cuts off LP solutons (Gomory cut) Heurstcs: fnd good ntal ncumbent soluton Parallel: use separate cores to solve nodes n B&B tree Speedup from 990 04: 0,000 computer speed 580,000 algorthm mprovements 09

MILP Formulaton of UFL mn j j N N jm s.t. x, jm where N my x, N ky 0 x, N, j M j jm j j y 0,, N cx k fxed cost of NF at ste N,..., n c varable cost from to serve EF j M,..., m j, f NF establshed at ste y 0, otherwse x fracton of EF j demand served from NF at ste. j y x, N, j M j 0

Capactated Faclty Locaton (CFL) mn ky j j N N jm s.t. x, jm N Ky fx, N 0 x, N, j M j j j jm j cx y 0,, N where k fxed cost of NF at ste N,..., n c varable cost from to serve EF j M,..., m j K capacty of NF at ste N,..., n f demand EF jm,..., m j, f NF establshed at ste y 0, otherwse xj fracton of EF j demand served from NF at ste. CFL does not have smple and effectve heurstcs, unlke UFL Other types of constrants: Fx NF at ste j: set LB and UB of x j to Convert UFL to p Medan: set all k to 0 and add constrant sum{y } = p

Matlog s Mlp Executng mp = Mlp creates a Mlp object that can be used to defne a MILP model that s then passed to a Solver Smlar syntax to math notaton for MILP AMPL and OPL algebrac modelng languages provde smlar capabltes, but Mlp ntegrated nto MATLAB

Ex: Illustrate Mlp syntax

Ex: UFL mn j j N N jm s.t. x, jm N my x, N ky 0 x, N, jm j jm j j y 0,, N cx 4

4 M M * (Weghted) Set Coverng,..., m, objects to be covered M M, N,..., n, subsets of M c cost of usng M n cover I arg mn c : M M, mn cost coverng of M I I I M 4 M M M 5 5 6 I M *,...,6 N,...,5,,, 4, 5,, 5,,6, M 6 M M M M 4 5 c, for all N * I arg mn c : M M c, 4 I I I 5

mn N (Weghted) Set Coverng s.t. a x, j M N M j x 0,, N * cx,..., m, objects to be covered M M, N,..., n, subsets of M c cost of usng M n cover I arg mn c : M M, mn cost coverng of M I I I where a x j, f M s n cover 0, otherwse, f j M 0, otherwse. 6

Set Packng Maxmze the number of mutually dsjont sets Dual of Set Coverng problem Not all objects requred n a packng Lmted logstcs engneerng applcaton (c.f. bn packng) M M M 5 max N s.t. a x, j M N j x 0,, N x 4 5 6 7

mn M y s.t. Vy v x, M where y x j j j jm M x, jm j y 0,, M x 0,, M, jm j M v j * Bn Packng,..., m, objects to be packed volume of object j V volume of each bn B max v V B arg mn B : v V, B M, mn packng of M B j jb BB, f bn B s used n packng 0, otherwse, f object j packed nto bn B 0, otherwse. 8 j