Presentation on theme: "Y. Pochet, Lhoist Group IP at CORE May 27-29, 2009"— Presentation transcript:
1 Y. Pochet, Lhoist GroupIP at COREMay 27-29, 2009From PW and MIP to PP by MIP From Production Waste and Milling Intermediate Products To Production Performance by Milling Inequality Polyhedron
2 Outline Supply Chain Network at Lhoist Lime Basics Production Waste Constraint and Model for the Milling ProcessMilling Inequality PolyhedronImproved Production Performance using the Milling Inequality Polyhedron
3 Supply Chain Network Optimization at Lhoist Lhoist GroupLhoist is a privately held lime companyHeadquarters in Limelette, Belgium, EuropeLhoist is the leading lime manufacturer in the world with over 70 operations in:Belgium, France, Germany, Denmark,Poland, Czech Republic, England, Spain, Portugal, North- and Central AmericaBrazilSupply Chain Network Optimization at LhoistLhoist Group is using an APS (Advanced Planning Software) as part of its DSS (Decision Support System) for its strategic and tactical “supply chain network optimizations”.Development started in 1991, research project Lhoist (P. Sevrin) – UCL (Y. Pochet)Specificity: divergent product structure with linked co-products (grain size, physical and chemical quality, variable recipes,…) difficult flow balance constraints within the plants.The system and the optimization approach have been used extensively within Lhoist Group for many strategic studies. The system will never be a “decision system” and will remain a support to challenge creativity and entrepreneurship.
7 Outline Supply Chain Network at Lhoist Lime Basics Production Waste Constraint and Model for the Milling ProcessMilling Inequality PolyhedronImproved Production Performance using the Milling Inequality Polyhedron
8 I don’t have the foggiest idea what this is about… OutlineSupply Chain Network at LhoistLime BasicsProduction Waste Constraint and Model for the Milling ProcessMilling Inequality PolyhedronImproved Production Performance using the Milling Inequality PolyhedronAs Karen would say:I don’t have the foggiest idea what this is about…Let’s try again!
9 This is not only about science This is about something else …
10 Y. Pochet, Lhoist GroupIP at COREMay 27-29, 2009From PW and MIP to PP by MIP From (Pochet-) Wolsey and Mixed Integer Programming To Production Planning by Mixed Integer Programming
11 The Goal: Thank you Laurence for your exceptional guidance! Laurence’s PhD Students (+ many others, sorry ! )Work inspired by LaurenceConstant drive to develop his studentsFascination for the Lot-Sizing worldArea/Era 1: The Happy few or PP and MIPSingle item planning modelsSpecific Reformulations (cutting planes, extended,…)Area/Era 2: The Happy many or PP by MIPMulti item production planning modelsOptimization algorithms & SystemsGeneralizations to MIPs & Generic reformulationsArea 3: The Happy all or MIP/IPFacility LocationScheduling and Constraint ProgrammingPartitioning ProblemsGraphs with Bounded Decomposability ; FlowsMarkov and Groebner Bases
12 Laurence’s PhD Students Y. Pochet, Lot-Sizing Problems: Reformulations and Cutting Plane Algorithms (1987)C. Bousba, Planification des Réseaux Electriques de Distribution à Basse Tension: une Approche par la Programmation Mathématique (1989)J.P. de Sousa, Time Indexed Formulations of Non-Preemptive Single-Machine Scheduling Problems (1989)K. Aardal, On the Solution of One and Two-Level Capacitated Facility Location Problems by the Cutting Plane Approach (1992)E-H Aghezzaf, Optimal Constrained Rooted Subtrees and Partitioning Problems on Tree Graphs (1992)C. de Souza, The Graph Equipartition Problem: Optimal Solutions, Extensions and Applications (1993)M. Schaffers, On Links between Graphs with Bounded Decomposability, Existence of Efficient Algorithms, and Existence of Polyhedral Characterizations (1994)F. Vanderbeck, Decomposition and Column Generation for Integer Programs (1994)M. Constantino, A Polyhedral Approach to Production Planning Models: Start-Up Costs and Times, Upper and Lower Bounds on Production (1995)
13 Laurence’s PhD Students H. Marchand, A Polyhedral Study of the Mixed Knapsack Set and its Use to Solve Mixed Integer Programs (1998)G. Belvaux, Modelling and Solving Lot-Sizing Problems by Mixed Integer Programming (1999)C. Cordier, Development and Experimentation with a Branch and Cut System for Mixed Integer Programming (1999)M. Loparic, Stronger Mixed 0-1 Models for Lot-Sizing Problems (2001)F. Ortega, Formulations and Algorithms for Fixed Charge Networks and Lot-Sizing Problems (2001)M. Van Vyve, A Solution Approach of Production Planning Problems based on Compact Formulations for Single-Item Lot-Sizing Models (2003)Q. Louveaux, Exploring Structure and Reformulations in Different Integer Programming Algorithms(2004)J-F. Macq, Optimization of Multimedia Flows over Data Networks (2005)R. Sadykov, Integer Programming-based Decomposition Approaches for Solving Machine Scheduling Problems (2006)P. Malkin, Computing Markov bases, Groebner bases and extreme rays (2007)
14 Area 1: Lot Sizing Models LS-C Single Item : LS-U ; LS-C ; LS-CC (Wagner-Whitin costs WW ; Discrete Prod. DLS) Variants : Backlogging [LS,WW,DLS]1 - [U,C,CC]1 / B Start-Up Costs [LS,WW,DLS]1 - [U,C,CC]1 / SC Start-Up Times [LS,WW,DLS]1 - [U,C,CC]1 / ST Sales (profit max) [LS,WW,DLS]1 - [U,C,CC]1 / SL Safety Stocks [LS,WW,DLS]1 - [U,C,CC]1 / SS Lower Bounds [LS,WW,DLS]1 - [U,C,CC]1 / LB Research Stream on Algorithms ; Valid Inequalities ; Extended Reformulations
17 Thank you so much Laurence for your Patience (personal comment)Permanent challenging mindset: Asking (us) the right - but tough - questions at the right timeConstant DriveIntuition and deep knowledge of the fieldContinuous source of inspiration and intellectual motivationGuidanceFriendship
18 Yves, Choaib, Jorge, Karen, El Houssaine, Cid, Michel, François, Miguel, Hugues,Gaëtan, Cécile, Marko,Francisco dit Pancho,Mathieu, Quentin,Jean-François, Ruslan, Peter.