Network Optimization Problems. Chapter 6: Hillier and Hillier

Similar documents
NETWORK OPTIMIZATION MODELS

Overview. H. R. Alvarez A., Ph. D.

Network Optimization Models. Chapter 10: Hillier and Lieberman Chapter 8: Decision Tools for Agribusiness Dr. Hurley s AGB 328 Course

1. What do you get as the integer and noninteger parts if you factor this as we did with our cutting planes:

Introduction to Management Science (8th Edition, Bernard W. Taylor III) Chapter 12 Network Flow Models. Chapter 12 - Network Flow Models 1

Linear Programming: Basic Concepts. Chapter 2: Hillier and Hillier

Lecture 8 Network Flow Models

Chapter 10: Network Flow Programming

IV. Special Linear Programming Models

Introduction to Mathematical Programming IE406. Lecture 16. Dr. Ted Ralphs

APPM 4120/5120 Exam #2 Practice Solutions Spring 2015

4. Linear Programming

Lecture notes on Transportation and Assignment Problem (BBE (H) QTM paper of Delhi University)

Network Flow Models. Chapter Copyright 2010 Pearson Education, Inc. Publishing as Prentice Hall

Notes on Network Optimization CE 4920 Stephen Boyles Fall 2009

Optimization Methods in Management Science

Ch.07 Shortest Route, Minimal Spanning Tree, and Maximal Flow Models. Management Science / Prof. Bonghyun Ahn

Unit.9 Integer Programming

Transportation Planning: The maximum flow problem

Prepared By. Handaru Jati, Ph.D. Universitas Negeri Yogyakarta.

Linear Programming Motivation

Application of Bounded Variable Simplex Algorithm in Solving Maximal Flow Model

Linear programming II João Carlos Lourenço

CSC 8301 Design & Analysis of Algorithms: Linear Programming

CSCI 5654 (Fall 2013): Network Simplex Algorithm for Transshipment Problems.

Copyright 2007 Pearson Addison-Wesley. All rights reserved. A. Levitin Introduction to the Design & Analysis of Algorithms, 2 nd ed., Ch.

2/25/2016. Minimal Spanning Tree Problem. Minimal Spanning Tree Problem. Example: Minimal Spanning Tree. Example: Minimal Spanning Tree

Introduction to Linear Programming. Chapter 3: Hillier and Lieberman Chapter 3: Decision Tools for Agribusiness Dr. Hurley s AGB 328 Course

ME 391Q Network Flow Programming

Fundamentals of Operations Research. Prof. G. Srinivasan. Department of Management Studies. Indian Institute of Technology, Madras. Lecture No.

Chalmers University of Technology. Network Models. April 11, Birgit Grohe

Network Models - Examples. Network Models. Definition and Terminology. The Minimum Spanning Tree (MST) Problem

ORF 307: Lecture 14. Linear Programming: Chapter 14: Network Flows: Algorithms

Network optimization: An overview

Dynamic modelling and network optimization

of optimization problems. In this chapter, it is explained that what network design

Chapter 5 Graph Algorithms Algorithm Theory WS 2012/13 Fabian Kuhn

The Islamic University of Gaza Faculty of Commerce Quantitative Analysis - Dr. Samir Safi Midterm #2-28/4/2014

Outline: Finish uncapacitated simplex method Negative cost cycle algorithm The max-flow problem Max-flow min-cut theorem

15.082J and 6.855J. Lagrangian Relaxation 2 Algorithms Application to LPs

TIM 206 Lecture Notes Integer Programming

ADMS 3330 FALL 2008 EXAM All Multiple choice Exam (See Answer Key on last page)

Material handling and Transportation in Logistics. Paolo Detti Dipartimento di Ingegneria dell Informazione e Scienze Matematiche Università di Siena

Mathematical Tools for Engineering and Management

Lesson 08 Linear Programming

COT 6936: Topics in Algorithms! Giri Narasimhan. ECS 254A / EC 2443; Phone: x3748

TRANSPORTATION AND ASSIGNMENT PROBLEMS

Transportation Problems

CSE 417 Network Flows (pt 4) Min Cost Flows

BCN Decision and Risk Analysis. Syed M. Ahmed, Ph.D.

Simulation. Lecture O1 Optimization: Linear Programming. Saeed Bastani April 2016

Lecture 2 Generalized Representation of a Transportation Network

What is Network Analyst?

We have already seen the transportation problem and the assignment problem. Let us take the transportation problem, first.

Solutions for Operations Research Final Exam

SOLVING THE TRANSPORTATION PROBLEM WITH PIECEWISE-LINEAR CONCAVE COST FUNCTIONS

= 5. x 3,1. x 1,4. x 3,7 =15. =10 x 6,3 = 8

MULTIMEDIA UNIVERSITY FACULTY OF ENGINEERING PEM2046 ENGINEERING MATHEMATICS IV TUTORIAL

Konigsberg Bridge Problem

A compromise method for solving fuzzy multi objective fixed charge transportation problem

AMPL Network Optimization

Foundations of Operations Research Introduction to AMPL and Modellisation

Graph Algorithms Maximum Flow Applications

Advanced Operations Research Prof. G. Srinivasan Department of Management Studies Indian Institute of Technology, Madras

OIM 413 Logistics and Transportation Lecture 2: Representation of a Transportation Network

Introduction. Chapter 15. Optimization Modeling: Applications. Integer Programming. Manufacturing Example. Three Types of ILP Models

Chapter 16. Greedy Algorithms

CMPSCI 311: Introduction to Algorithms Practice Final Exam

INEN 420 Final Review

Final Exam Spring 2003

PBW 654 Applied Statistics - I Urban Operations Research. Unit 3. Network Modelling

Advanced Operations Research Prof. G. Srinivasan Department of Management Studies Indian Institute of Technology, Madras

Network Analysis. Links, nodes, trees, graphs, paths and cycles what does it all mean? Minimal spanning tree shortest route maximum flow

Fundamentals of Integer Programming

Approximation Algorithms

CSE 40/60236 Sam Bailey

Computational Methods in IS Research Fall Graph Algorithms Network Flow Problems

CSE 417 Network Flows (pt 3) Modeling with Min Cuts

Fundamentals of Operations Research. Prof. G. Srinivasan. Department of Management Studies. Indian Institute of Technology Madras.

Greedy Homework Problems

Primal Simplex Algorithm for the Pure Minimal Cost Flow Problem

Integer Programming. Xi Chen. Department of Management Science and Engineering International Business School Beijing Foreign Studies University

A Computer Oriented Method for Solving Transportation Problem

OPERATIONS RESEARCH. Linear Programming Problem

Linear Programming. L.W. Dasanayake Department of Economics University of Kelaniya

Department of Mathematics Oleg Burdakov of 30 October Consider the following linear programming problem (LP):

18 Spanning Tree Algorithms

Using the Graphical Method to Solve Linear Programs J. Reeb and S. Leavengood

Problem set 2. Problem 1. Problem 2. Problem 3. CS261, Winter Instructor: Ashish Goel.

Using the Simplex Method to Solve Linear Programming Maximization Problems J. Reeb and S. Leavengood

The University of Sydney MATH 2009

Chapter 1 Introduction to Optimization

The k-center problem Approximation Algorithms 2009 Petros Potikas

Linear Programming. Readings: Read text section 11.6, and sections 1 and 2 of Tom Ferguson s notes (see course homepage).

OPTIMIZAÇÃO E DECISÃO 09/10

Linear Programming. Widget Factory Example. Linear Programming: Standard Form. Widget Factory Example: Continued.

Linear Programming CISC4080, Computer Algorithms CIS, Fordham Univ. Linear Programming

4 LINEAR PROGRAMMING (LP) E. Amaldi Fondamenti di R.O. Politecnico di Milano 1

Artificial Intelligence

CMPSCI611: The Simplex Algorithm Lecture 24

Transcription:

Network Optimization Problems Chapter 6: Hillier and Hillier

Agenda Minimum-Cost Flow Problems Case Study: Distribution Unlimited Co. Maximum Flow Problems Case Study: BMZ Company Shortest Path Problems Case Study: Littletown Fire Department Minimum Spanning Tree Problems Case Study: The Modern Corp. Problem

Minimum-Cost Flow Problems These problems are also known as distribution network problems which were seen in chapter 3. Your objective is to distribute goods from a supply source to an ultimate demand destination with potential distribution centers in-between.

Needed Terminology A network in this case is a representation of the model for a minimum-cost flow model that has flow passing through it. Circles in the network are known as nodes. These nodes can be broken-up into supply nodes, demand nodes, and transshipment nodes. A supply node is a node where the net amount of flow generated (outflow minus inflow) is a fixed positive number.

Needed Terminology Cont. A demand node is a node where the net amount of flow generated (outflow minus inflow) is a fixed negative number. A transshipment node is a node where the net amount of flow generated (outflow minus inflow) is a fixed at zero. Having the outflow equal to the inflow is known as conservation of flow.

Needed Terminology Cont. Arcs are lines from node to node which represent the direction in which flow can occur. Capacity of an arc represents the maximum amount of flow that can be sent through an arc. Flow is defined as an amount that is sent through an arc.

Assumptions of the Minimum- Cost Flow Problem There is at least one supply node. There is at least one demand node. Any node that is not a supply node or a demand node is a transshipment node. Flow through an arc is only allowed in the direction the arc is pointing, where the maximum amount of flow is dictated by its capacity.

Assumptions of the Minimum- Cost Flow Problem The network has enough arcs with sufficient capacity to move all the flows from the supply nodes to the demand nodes. The cost of the flows through the arc are proportional to the amounts sent through the arc. The cost of each flow is known. The objective is to minimize the cost of the flows from the supply nodes to the demand nodes.

Feasible Solution and Integer Properties Feasible Solution Property A minimum-cost flow problem is said to have a feasible solution if and only if the sum of the supplies is equal to the sum of the demands. Integer Solution Property A minimum-cost flow problem will have an integer solution as long as all its supplies, demands, and arc capacities have integer values.

Case Study: Distribution Unlimited Co. Distribution Unlimited Co. (DUC) has two factories producing a product that needs to be shipped to two warehouses. Factory 1 produces 80 units Factory 2 produces 70 units Warehouse 1 needs 60 units Warehouse 2 needs 90 units DUC has one distribution center that both factories can ship to.

Case Study: Distribution Unlimited Co. Cont. DUC has the ability for factory 1 to ship to warehouse 1 or the distribution center. DUC has the ability for factory 2 to ship to warehouse 2 or the distribution center. The distribution center can ship to either warehouses. Shipments to and from distribution centers are limited to a 50 unit maximum amount for each route.

Case Study: Distribution Unlimited Co. Cont. Unit Cost For Shipments Distribution Warehouse 1 Warehouse 2 Center Factory 1 $300 $700 N/A Factory 2 $400 N/A $900 Distribution Center N/A $200 $400

Case Study: Distribution Unlimited Co. Cont. Maximum Amount of Shipments to Each Source Distribution Warehouse 1 Warehouse 2 Center Factory 1 50 50 N/A Factory 2 50 N/A 50 Distribution Center N/A 50 50

Case Study: Distribution Unlimited Co. Cont. DUC wants to find the cheapest shipping plan to transport is supply from the factories to the demand of the warehouses.

Representation of Distribution Unlimited Co. Problem 80 units produced F1 W1 60 units needed DC 70 units produced F2 W2 90 units needed

Representation of Distribution Unlimited Co. Problem Cont. [80] [- 60] F1 $700 W1 $300 $200 [50] [0] [50] DC F2 $400 [50] $900 $400 [50] W2 [70] [- 90]

Mathematical Model of Distribution Unlimited Co. 2 400* 1 200* 2 2 900* 2 400* 1 300* 1 1 700* 2 1, 2 2, 2 1 1, 1 DCW DCW W F DC F DC F W F Min DCW DCW W F DC F DC F W F + + + + +

Mathematical Model of Distribution Unlimited Co. Cont. Constraints : F1DC + F2DC - DCW1- DCW2 = 0 F1W1+ DCW1 = 60 F2W2 + DCW2 = 90 F1W1+ F1DC = 80 F2W2 + F2DC = 70 F1DC, F2DC, DCW1, DCW2 50 F1DC, F2DC, DCW1, DCW2, F1W1, F2W2 0

SUMIF Function in Excel The SUMIF function takes a range and compares it to a some condition. If the condition is true, then the function will add to the running sum the corresponding value in another range. If the condition is false, then the function will not add to the running sum the corresponding value in another range. SUMIF(Range A, condition, Range B) If you put in a value or a number, the function checks for equality.

A Note on Solutions While the Simplex Method can solve minimum-cost flow problems, it is not the most efficient algorithm to use. The most efficient algorithm is the the Network Simplex Method. Unfortunately with Excel s Solver, the network Simplex cannot be utilized because Solver does not currently have the algorithm.

Maximum Flow Problems The objective of a maximum flow problem is to find a plan that maximizes the flow through the network. The characteristics of the maximum flow problem are similar to those of the minimum-cost flow problem.

Assumptions of the Maximum Flow Problem All flow originates from a single node known as the source. All flow terminates at a single node known as the sink. All remaining nodes are transshipment nodes. Flow through the arc is only allowed by the direction of the flow arrow.

Assumptions of the Maximum Flow Problem Maximum flow through an arc is dictated by the capacity of the arc. All arcs extending from the source node point away from the source node. All arcs coming to the sink node point towards the sink node. The objective is to maximize the total flow from the source to the sink. The total flow can be either measured by the amount entering the sink or the amount leaving the source.

Differences Between Minimum- Cost and Maximum Flow In the minimum-cost model, the supply nodes and demand nodes have fixed amounts, whereas the maximum flow problem has a variable amount. In the minimum-cost model, there can be multiple demand and supply nodes, whereas the maximum flow model has only one supply node and one demand node.

Case Study: BMZ Company BMZ is a car company that is trying to get as many cars from its factory in Stuttgart to the Los Angeles market. BMZ can send cars across rail to three shipping points to the US: Rotterdam Bordeaux Lisbon

Case Study: BMZ Company Cont. From the Rotterdam terminal, cars can be shipped to a port in New York. From Bordeaux, cars can be shipped to either New York or New Orleans. From Lisbon, cars can be shipped to New Orleans. Once the cars are in New York and New Orleans, the cars will be shipped to the LA market. Each route for shipment has a limited amount of cars that it can handle.

Case Study: BMZ Company Cont. Maximum Shipments Allowed Rotterdam Bordeaux Lisbon New York New Orleans Los Angeles Stuttgart 50 70 40 N/A N/A N/A Rotterdam N/A N/A N/A 60 N/A N/A Bordeaux N/A N/A N/A 40 50 N/A Lisbon N/A N/A N/A N/A 30 N/A New York N/A N/A N/A N/A N/A 80 New Orleans N/A N/A N/A N/A N/A 70

BMZ Company Visual Diagram [60 units max.] RO Rotterdam New York [80 units max.] NY [50 units max.] {40 units max.] BO ST [70 units max.] Stuttgart Bordeaux [40 units max.] [50 units max.] LA New Orleans Los Angeles [70 units max] NO [30 units max.] LI Lisbon

Mathematical Formulation 70 80, 30, 50, 40, 60, 40, 70, 50, STRO Constraints :,,,,,,,, + NOLA NYLA LINO BONO BONY RONY STLI STBO NOLA NYLA Max NOLA NYLA LINO BONO BONY RONY STLI STBO STRO

Mathematical Formulation Cont. 0 0,,, 0,,, STRO, 0 0 0 0 0 STRO Constraints Cont.: = + = + = + = = = NOLA LINO BONO NOLA NYLA LINO STLI BONO BONY STBO RONY NOLA LINO BONO NYLA RONY BONY LINO STLI BONO BONY STBO RONY

Shortest Path Problems In the shortest path problem the objective is to find the shortest path from an origin node to a destination node.

Assumptions of the Shortest Path Problem You need to choose a path that starts at an origin node and ends at a destination node and follows a particular path. The key to these types of problems is that ordering of the solution matters. The lines connecting nodes in the shortest path problem are known as links and usually are bidirectional. Arcs are also allowed.

Assumptions of the Shortest Path Problem Cont. Associated with each link is a nonnegative number known as the length. Lengths can be considered time, costs, distance, etc. The objective is to find the shortest path from the origin node to the destination node.

Case Study: Littletown Fire Department Littletown Fire Department is trying to find the fastest route through the county to reach a specific farming community. The fire department would like to know what is the shortest path to get from the fire department to the farming community.

Diagram of Littletown s Problem 8 6 A F 4 Fire Station 4 3 6 C 1 B 2 4 5 7 D 3 E 4 6 3 5 H G 2 6 7 Farming Community

Minimum Spanning Tree Problems (Not in textbook) Minimum spanning tree problems are problems where you are trying to find the most efficient network. You start out with nodes that have potential links and you must decide whether a link from two nodes should exist. The objective in spanning tree problems are to link all the nodes directly or indirectly to each other with the smallest amount of resources used.

Needed Terminology Tree A tree is a network that does not have any paths that begin and end at the same node. Spanning Tree It is a tree that provides a path between every pair of nodes, i.e., it spans the tree.

Assumptions of the Minimum Spanning Tree Problem You are given the nodes of the network, but not the links. You are given all the costs related to linking nodes together. You must put enough links into the network to link every pair of nodes so that the nodes are either directly or indirectly linked. The objective is to choose the links that satisfy the above at the cheapest cost.

Algorithm for a Minimum Spanning Tree Problem The Greedy Algorithm Step 1: Select the cheapest potential link. Step 2: Select the cheapest potential link between a node that already has a link with one that does not have a link. Step 3: Continue step two until all nodes have been directly or indirectly linked. Note: Ties may be broken arbitrarily.

Notes on Minimum Spanning Trees The number of links in a spanning tree are one less than the number of nodes. Each node is directly connected by a single link to at least one other node. It is possible to have multiple minimum spanning trees.

Case Study: The Modern Corp. Problem Modern Corp is trying to provide fiber-optic network to all its major centers. The company only needs to make sure that the centers are directly or indirectly linked to each other. They want to link the centers with the cheapest amount of links. The potential links are given on the next slide.

The Modern Corp. Problem A B 2 2 5 C 7 4 E 5 G 4 1 3 1 7 D 4 F