Amiri s Supply Chain Model. System Engineering b Department of Mathematics and Statistics c Odette School of Business

Similar documents
The Minimum Universal Cost Flow in an Infeasible Flow Network

Integrated approach in solving parallel machine scheduling and location (ScheLoc) problem

Winter 2008 CS567 Stochastic Linear/Integer Programming Guest Lecturer: Xu, Huan

Single-Facility Scheduling over Long Time Horizons by Logic-based Benders Decomposition

Solution (1) Formulate the problem as a LP model.

Simultaneous Optimization of Berth Allocation, Quay Crane Assignment and Quay Crane Scheduling Problems in Container Terminals

A Modified Vogel s Approximation Method for Obtaining a Good Primal Solution of Transportation Problems

An Admission Control Algorithm in Cloud Computing Systems

SOLVING CAPACITATED VEHICLE ROUTING PROBLEMS WITH TIME WINDOWS BY GOAL PROGRAMMING APPROACH

En Route Traffic Optimization to Reduce Environmental Impact

A PROBABILITY-DRIVEN SEARCH ALGORITHM FOR SOLVING MULTI-OBJECTIVE OPTIMIZATION PROBLEMS

A Simple Inventory System

An Interactive Optimisation Tool for Allocation Problems

Interactive Bi-Level Multi-Objective Integer. Non-linear Programming Problem

THE DETERMINATION OF PARADOXICAL PAIRS IN A LINEAR TRANSPORTATION PROBLEM

Solutions to exam in SF1811 Optimization, Jan 14, 2015

Module 3 LOSSY IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur

EEL 6266 Power System Operation and Control. Chapter 3 Economic Dispatch Using Dynamic Programming

ECE559VV Project Report

Outline and Reading. Dynamic Programming. Dynamic Programming revealed. Computing Fibonacci. The General Dynamic Programming Technique

An Effective Modification to Solve Transportation Problems: A Cost Minimization Approach

The Value of Demand Postponement under Demand Uncertainty

On the Multicriteria Integer Network Flow Problem

Chapter - 2. Distribution System Power Flow Analysis

SOLVING MULTI-OBJECTIVE INTERVAL TRANSPORTATION PROBLEM USING GREY SITUATION DECISION-MAKING THEORY BASED ON GREY NUMBERS

COS 521: Advanced Algorithms Game Theory and Linear Programming

Heuristic Algorithm for Finding Sensitivity Analysis in Interval Solid Transportation Problems

CIE4801 Transportation and spatial modelling Trip distribution

Natural Language Processing and Information Retrieval

Assortment Optimization under MNL

Dynamic Facility Location with Stochastic Demands

FUZZY GOAL PROGRAMMING VS ORDINARY FUZZY PROGRAMMING APPROACH FOR MULTI OBJECTIVE PROGRAMMING PROBLEM

Journal of Physics: Conference Series. Related content PAPER OPEN ACCESS

TRAPEZOIDAL FUZZY NUMBERS FOR THE TRANSPORTATION PROBLEM. Abstract

Resource Allocation with a Budget Constraint for Computing Independent Tasks in the Cloud

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur

Module 9. Lecture 6. Duality in Assignment Problems

Dynamic Programming. Lecture 13 (5/31/2017)

See Book Chapter 11 2 nd Edition (Chapter 10 1 st Edition)

An Integrated OR/CP Method for Planning and Scheduling

LP-based Approximation Algorithms for Capacitated Facility Location

MMA and GCMMA two methods for nonlinear optimization

LINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity

A Lower Bound on SINR Threshold for Call Admission Control in Multiple-Class CDMA Systems with Imperfect Power-Control

A LINEAR PROGRAM TO COMPARE MULTIPLE GROSS CREDIT LOSS FORECASTS. Dr. Derald E. Wentzien, Wesley College, (302) ,

10.34 Numerical Methods Applied to Chemical Engineering Fall Homework #3: Systems of Nonlinear Equations and Optimization

Technical Note: Capacity Constraints Across Nests in Assortment Optimization Under the Nested Logit Model

College of Computer & Information Science Fall 2009 Northeastern University 20 October 2009

IJRSS Volume 2, Issue 2 ISSN:

Computing Correlated Equilibria in Multi-Player Games

Fuzzy approach to solve multi-objective capacitated transportation problem

Department of Statistics University of Toronto STA305H1S / 1004 HS Design and Analysis of Experiments Term Test - Winter Solution

Lecture 10 Support Vector Machines II

STAT 511 FINAL EXAM NAME Spring 2001

Formulations, Branch-and-Cut and a Hybrid Heuristic Algorithm for an Inventory Routing Problem with Perishable Products

Dual-Channel Warehouse and Inventory Management with Stochastic Demand

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

Predictive Analytics : QM901.1x Prof U Dinesh Kumar, IIMB. All Rights Reserved, Indian Institute of Management Bangalore

Solution Thermodynamics

Structure and Drive Paul A. Jensen Copyright July 20, 2003

Supply Network Protection under Capacity Constraint

COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS

Assignment 4. Adsorption Isotherms

Chapter Newton s Method

Maximizing the number of nonnegative subsets

Hierarchical State Estimation Using Phasor Measurement Units

An Upper Bound on SINR Threshold for Call Admission Control in Multiple-Class CDMA Systems with Imperfect Power-Control

Calculation of time complexity (3%)

Journal of Applied Science and Agriculture

Dynamic Slope Scaling Procedure to solve. Stochastic Integer Programming Problem

CHAPTER 7 STOCHASTIC ECONOMIC EMISSION DISPATCH-MODELED USING WEIGHTING METHOD

The written Master s Examination

Finding Optimal System Time for Message Processing in Free-Sequence Multi-Stage News Agency

Which Separator? Spring 1

Using T.O.M to Estimate Parameter of distributions that have not Single Exponential Family

Absorbing Markov Chain Models to Determine Optimum Process Target Levels in Production Systems with Rework and Scrapping

Polynomial Regression Models

LATERAL TRANSSHIPMENTS IN INVENTORY MODELS

find (x): given element x, return the canonical element of the set containing x;

Open Systems: Chemical Potential and Partial Molar Quantities Chemical Potential

MEM Chapter 4b. LMI Lab commands

Chapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems

princeton univ. F 17 cos 521: Advanced Algorithm Design Lecture 7: LP Duality Lecturer: Matt Weinberg

Basic Statistical Analysis and Yield Calculations

Presenters. Muscle Modeling. John Rasmussen (Presenter) Arne Kiis (Host) The web cast will begin in a few minutes.

OPTIMISATION. Introduction Single Variable Unconstrained Optimisation Multivariable Unconstrained Optimisation Linear Programming

How Strong Are Weak Patents? Joseph Farrell and Carl Shapiro. Supplementary Material Licensing Probabilistic Patents to Cournot Oligopolists *

INTEGER NETWORK SYNTHESIS PROBLEM FOR HOP CONSTRAINED FLOWS

Extensions of the Psuedo-Boolean Representation of the p-median Problem

Multi-product budget-constrained acquistion and pricing with uncertain demand and supplier quantity discounts

Siqian Shen. Department of Industrial and Operations Engineering University of Michigan, Ann Arbor, MI 48109,

Service-Level Differentiation in Many-Server Service Systems via Queue-Ratio Routing

An Integrated Strategy for a Production Planning and Warehouse Layout Problem: Modeling and Solution Approaches

Supplementary Notes for Chapter 9 Mixture Thermodynamics

Chapter 3 Describing Data Using Numerical Measures

For now, let us focus on a specific model of neurons. These are simplified from reality but can achieve remarkable results.

THE SUMMATION NOTATION Ʃ

NP-Completeness : Proofs

8. Modelling Uncertainty

Some modelling aspects for the Matlab implementation of MMA

Transcription:

Amr s Supply Chan Model by S. Ashtab a,, R.J. Caron b E. Selvarajah c a Department of Industral Manufacturng System Engneerng b Department of Mathematcs Statstcs c Odette School of Busness Unversty of Wndsor Wndsor, ON N9B P4 WMSR # 14-01

Amr s Supply Chan Model Sah Ashtab a,, Rchard J. Caron b, Esagnan Selvarajah c a Department of Industral Manufacturng Systems Engneerng, Unversty of Wndsor, Wndsor, ON, Canada, N9B P4 b Department of Mathematcs Statstcs, Unversty of Wndsor, Wndsor, ON, Canada, N9B P4 c Odette School of Busness, Unversty of Wndsor, Wndsor, ON, Canada, N9B P4 Abstract We show that Amr s model admts alternate optmal solutons that are nconsstent wth the model s phlosophy we provde a set of constrants that ensure consstency of all optmal solutons. Keywords: Supply chan network desgn, Faclty locaton, Mult-echelon, Mult-capactated. 1. Introducton Amr [1] presented a model Lagrangan-based soluton technque for mult-echelon, mult-capactated supply chan network desgn. We show that Amr s model admts alternate optmal solutons some of whch are nconsstent wth the model phlosophy. We provde a new set of constrants to ensure that the optmal soluton provded by the model properly lnks the decson to buld a warehouse to the decsons to allocate dem to that warehouse; that the fractonal varables are ndeed, fractons, as requred. 2. The Amr Model [1] The set of customer zones s ndexed by, the potental warehouse locatons by j, the potental plant locatons by k, the warehouse capacty levels by r the plant capacty levels by h. Let a be the dem from customer zone durng the plannng horzon, b r j be the capacty of a warehouse at locaton j bult at level r e h k be the capacty of a plant at locaton k bult at level h. Let Yjk r be the fracton of the total dem of a warehouse at locaton j wth capacty level r that s delvered from a plant at locaton k let X j be the fracton of the total dem of customer zone that s delvered from a warehouse at locaton j. A warehouse wth capacty level r s bult at locaton Correspondng Author. Tel: +1-519-992-6274. Emal address: ashtab@uwndsor.ca (Sah Ashtab) Preprnt submtted to European Journal of Operatonal Research August 18, 2014

j f only f Uj r = 1 Vk h = 1 f only f a plant wth capacty level h s bult at locaton k. The bnary varables U V determne number, locaton capacty of the facltes whle the real-valued varables X Y determne the flow of goods. The objectve functon s to mnmze the sum of the transportaton costs the fxed costs of establshng plants warehouses. Let C jk be the cost of shppng one unt of dem to a warehouse at locaton j from a plant at locaton k let C j be the cost of shppng one unt of dem to customer zone from a warehouse at locaton j. The n-bound out-bound transportaton costs are gven by T 1 (Y ) = r,j,k C jk b r j Y r jk T 2 (X) =,j C j a X j, respectvely. The fxed costs of establshng operatng the warehouses plants over the plannng horzon are F 1 (U) = j,r F r j U r j F 2 (V ) = k,h G h k V h k, respectvely, where Fj r s the fxed cost of openng operatng warehouse at locaton j wth capacty level r G h k s the fxed cost of openng operatng plant at locaton k wth capacty level h. The objectve functon s a summaton of the four cost elements s gven by The constrants Z 0 (Y, X, U, V ) = T 1 (Y ) + T 2 (X) + F 1 (U) + F 2 (V ). X j = 1,, (1) j ensure that the dem at each customer zone s covered by bult warehouses whle constrants a X j b r j Uj r, j, (2) r ensure that the capacty level at each warehouse s suffcent to meet out-bound shpments. That each warehouse each plant s assgned a sngle capacty level s ensured by 1, j, () r h U r j respectvely. The set of constrants a X j k,r V h k 1, k, (4) b r j Y jk r, j, (5) 2

ensure that the total out-bound shpment from a warehouse s not greater than the total n-bound shpment from plants to that warehouse. The nequaltes j,r b r j Y r jk h e h k V h k, k, (6) ensure that the total n-bound shpment from a bult plant to the warehouses s not greater than the chosen capacty level of that plant. The fnal sets of constrants are X j 0,, j, (7) Y r jk 0, k, j, r, (8) whch, together wth the unstated constrants that Uj r V k h are bnary, gve Amr s model whch s to AM0: mn U,Vbnary { Z 0(Y, X, U, V ) (1) (8) }. The fact that there are alternate solutons, becomes clear wth the change of varables W jk = r b r jy r jk. (9) We now replace T 1 (Y ) wth T 1 (W) = j,k C jk W jk gvng the new expresson for the objectve Z 1 (W, X, U, V ) = T 1 (W) + T 2 (X) + F 1 (U) + F 1 (V ). We next replace replace constrants (5) (6) wth a X j W jk, j, (10) k W jk j h respectvely. Model AM0 s equvalent to e h k V h k, k, (11) AM1: mn U,Vbnary { Z 1(W, X, U, V ) (1) (4), (7), (8), (10), (11) } Gven the W jk from a soluton to AM1, any set of non-negatve solutons Y r jk to (9) s a soluton to AM0.

Example. Consder a supply chan wth one plant havng a sngle avalable capacty level of e 1 1 = 6 000 one warehouse wth avalable capacty levels of b 1 1 = 1 000, b 2 1 = 000 b 1 = 5 000. The three customer zones have dems a 1 = 1 500, a 2 = 1 500 a = 1 000. To be feasble we must have V1 1 = 1, U1 1 = 0, U1 2 = 0, U1 = 1, X 11 = 1, X 21 = 1, X 1 = 1 4 000 W 11 6 000. An optmal soluton wll have W 11 = 4 000. Consequently, any non-negatve soluton to 4 000 = 1 000 Y11 1 + 000 Y11 2 + 5 000 Y11 wll be feasble optmal for AM0. So, whle Y11 1 = 4, Y11 2 = 0, Y11 = 0 s optmal for AM0, t has Y varables greater than one whch s nconsstent wth the model phlosophy that they are fractons. The soluton Y11 1 = 1, Y11 2 = 1, Y11 = 0 s optmal for AM0 but t s nconsstent wth U1 1 = 0 U1 2 = 0. The example shows that the model allows alternate optmal solutons that could lead to two dfferent undesrable stuatons; Y varables greater than one Y varables nconsstent wth the U varables. Both of these stuatons are corrected wth the addton to the model of the constrants Yjk r Uj r, j, r. (12) The modfed model, wthout the change of varables, s k AM2: mn{ Z 0 (Y, X, U, V ) (1) (8), (12) }. For our example, these new constrant force the unque, acceptable, optmal soluton wth Y11 1 = 0, Y11 2 = 0, Y11 = 1. To test the mpact of ths addtonal set of constrants we created, usng the method descrbed by Amr, sx nstances of an example wth 500 customer zones, 0 warehouses 20 plants (the largest nstance solved by Amr) solved the AM0 the AM2 models usng LINGO 14.0 x64 on a DELL server wth two 2500 MHz CPUs. The tmes taken to reach optmalty for the AM0 ranged from 56 to 185 seconds wth an average of 122. The tmes taken to reach optmalty for the AM2 ranged from 5 to 277 seconds wth an average of about 121. Ths lmted numercal evdence suggests that the ablty to solve the model, n reasonable tme, s not compromsed by the addtonal constrants. References [1] A. Amr, Desgnng a dstrbuton network n a supply chan system: Formulaton effcent soluton procedure, European Journal of Operatonal Research 171 (2006) 567 576. 4