PTS312_6x - The Client Just-In-Time Scheduling: A Case...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: The Client Just-In-Time Scheduling: A Case Study in Operations Research Peter L. Jackson Professor School of Operations Research and Information Engineering April, 2007 Production-Transportation Scheduling 2 Client’s Statement of Problem • “They are having scheduling problems in Russia, Ohio. Everyone wants me to build them a finished goods warehouse. If I build them a warehouse, they will fill it full of inventory. Then, instead of a scheduling problem, I will have an inventory problem.” Vice President of Operations Production-Transportation Scheduling 3 Overview • • • • • • Production Highlights Transportation Highlights Proposed System Optimization Decision Support System Design Principles Production-Transportation Scheduling 4 Process Flow EPS Outer Roll Mill EPS Line St. Lawrence Press North Automatic Line Reed City Press Residential Roll Former Helm Press Steward and Blue Lines Shrink Wrap Tunnel Glazing #6 and #7 South Automatic Line Independent Line EPS Wrap EPS Glazing Wrap Station Pan Ruvo A Factory Tour EPS Inner Roll Mill Commercial Roll Former Tranemco Press Cut Shape Attach Glaze Wrap Production-Transportation Scheduling 5 Production-Transportation Scheduling 6 Staging Flags for Shipment Loading Docks Flag Storage Area Production-Transportation Scheduling 7 Data Collection • Start/Finish times by job for several days • (from previous industrial engineering Production Time 6000 5000 4000 CumRelease CumComplete study) 3000 • Typical routings • Estimated run rates per machine 2000 1000 0 0:00:00 4:48:00 9:36:00 14:24:00 19:12:00 0:00:00 Production-Transportation Scheduling 9 Production-Transportation Scheduling 10 Completion Delays Cumulative Releases with Completion Times 8000 7000 Cum. Panels Flag Completion Delays Cumulative Releases with Order Completion and Flag Completion Times CumRelease CumComplete CumShipment Cum. Panels 6000 5000 4000 3000 2000 1000 0 2/2/199 2/3/199 2/3/199 2/4/199 2/4/199 2/5/199 2/5/199 2/6/199 7 12:00 7 0:00 7 12:00 7 0:00 7 12:00 7 0:00 7 12:00 7 0:00 Date/Tim e CumRelease CumComplete 8000 6000 4000 2000 0 2/2/1 2/3/1 2/4/1 2/5/1 2/6/1 2/7/1 2/8/1 997 997 997 997 997 997 997 0:00 0:00 0:00 0:00 0:00 0:00 0:00 Date/Tim e Production-Transportation Scheduling 11 Production-Transportation Scheduling 12 Production Scheduling Overview • • • • • • Production Highlights Transportation Highlights Proposed System Optimization Decision Support System Design Principles Production-Transportation Scheduling 13 Production-Transportation Scheduling 14 Data Collection • Order history for 8 months • Customer, address (with zipcode) • Date ordered, date shipped • Flag (truckload ID), Carrier name • Stop number of drop on flag • For each order line item: • Model number, dimensions, quantity • Production line and routing details Quantity Shipped by Location } Production-Transportation Scheduling 15 Production-Transportation Scheduling 16 Open Orders Pareto Analysis Pareto Analysis of Zip Code Shipments 100% 90% 80% % of Cum Cost 70% 60% 50% 40% 30% 20% 10% 0% 0% 20% 40% 60% 80% 100% % of ZipCodes 2637 Zip Codes Production-Transportation Scheduling 17 Production-Transportation Scheduling 18 Drop Size Distribution Distribution of Drop Sizes 4500 4000 Number of Drops 3500 3000 2500 2000 1500 1000 500 0 0 20 40 Drop Size 60 80 100 Cumulative Flow Cumulative Orders and Shipments: April-Oct. 1996 600000 500000 Average Lead Time: 7.3 Days 400000 Cum. Order Cum. Ship 300000 200000 100000 8 Day Lead 0 01-Jun-96 08-Jun-96 16-Jun-96 25-Jun-96 03-Jul-96 11-Jul-96 18-Jul-96 26-Jul-96 02-Aug-96 09-Aug-96 16-Aug-96 23-Aug-96 31-Aug-96 09-Sep-96 16-Sep-96 01-May-96 09-May-96 17-May-96 24-May-96 24-Sep-96 08-Apr-96 16-Apr-96 30-Mar-96 24-Apr-96 02-Oct-96 10-Oct-96 18-Oct-96 24080 Drops Production-Transportation Scheduling 19 Production-Transportation Scheduling 20 Lead Time Performance Lead Time Performance 35000 30000 Lead Time Performance Lead Time Distribution 1 0.9 Cum. Relative Frequency 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 10 12 14 16 18 20 22 24 26 28 30 32 34 0 2 4 6 8 Number of Order Lines 25000 20000 15000 10000 5000 0 0 10 20 30 40 50 60 70 80 90 100 Lead Time (Days) Lead Tim e (Days) Production-Transportation Scheduling 21 Production-Transportation Scheduling 22 Vehicle Loading Doors Per Flag Distribution 70 60 Number of Flags 50 40 30 20 10 0 0 200 400 600 Doors per Flag 800 1000 1200 April 1996 Production-Transportation Scheduling 23 Production-Transportation Scheduling 24 Summary of Opportunities • Factory and distribution system design issues • Routing and sequencing problems in factory • Production loading problems • Vehicle loading, routing, and carrier choice problems in transportation Production-Transportation Scheduling 25 Overview • • • • • • Production Highlights Transportation Highlights Proposed System Optimization Decision Support System Design Principles Production-Transportation Scheduling 26 Enterprise Data Modeling Marketing Information System Purchasing Production Program Planning Sales/Order Processing Operations Sales/Dispatch Carrier Choice Flag Itinerary Flag Formation Shipping Schedule Raw Materials Allocation Manpower Schedule Production Schedule Workcenter Schedule P.M. Schedule Production Order Routing Choice Alternative Routings Shipping Units Freight List Sales/Order Processing Customer Open Order Accounts Payable Material Flow Control Personnel Information System Payroll Accounting Requirements Planning Sales/Dispatch Capacity Management/ Production Control Financial Accounting Accounts Receivable Material Flow Control Personnel Information System CAD Material Flow Control Office Automation A.W. Scheer, 1989 Material Flow Control Cost Accounting Capacity Management/ Production Control Decision Support Carrier Choice Overview Interactive Execute Display Edit PTS Stop Sequence Orders to Flags Flags to Days Flags to Shifts Operation Sequence Routing Choice PTS Sim Execute Interactive Display Edit • • • • • • Production Highlights Transportation Highlights Proposed System Optimization Decision Support System Design Principles Production-Transportation Scheduling 29 Production-Transportation Scheduling 30 Decomposition Ideas • Cluster first, route second • Orders to days using MILP Subsequent Ph.D. Dissertation After Trade Study Download Current Orders A1 Current Orders and Line Detail Scheduling Cycle A2 PTS Database • Orders to flags using greedy VRP User-Maintained Tables (Flag names, Carrier names, Pooling zipcodes) Build PTS Database Carrier Choice Stop Sequence Unscheduled Flags Form Loads and Sequence Stops A3 Order and Line Detail Views, Selection Queries, Hilighting and Subselections Map View, Map Selection Operations Stop Sequence Optimization Flag Formation Optimization Orders to Flags • Route first, cluster second • Orders to flags using greedy VRP • Flags to days using MILP • (Reoptimize: Orders to flags) Assign Flags to Production Days A4 Scheduled Flags Unscheduled Flag View, Scheduled Flag View, Drag and Drop Summary Totals View Export Revised Order File A5 Production Smoothing Optimization Revised Order File Upload Revised Orders A6 Production-Transportation Scheduling 31 Stop Sequence Optimization • Choices • Farthest insertion with two-opt • Farthest insertion with two-opt and Stop Sequence Stop Sequence Optimization Stop Sequence optional restrictions • TransCAD TSP algorithm • Variations on sweep algorithm (for very 84 LUMBER COMPANY #1220 NORTHGATE ELECTRIC DOORS INC C&S BUILDERS SUPPLY #1006800 ATLAS DOOR CORP large TSPs) RCS SOUTHSIDE • Implemented all of them and used rules/user choice to select Production-Transportation Scheduling 33 LOWES #683 ROSE CITY GARAGE DOORS INC Production-Transportation Scheduling 34 Stop Sequence Optimization Stop Sequence Flag Formation Optimization Selected Orders Form Single Flag F1 Initial Flag Carrier Choice Orders to Flags Initial Set of Flags 84 LUMBER COMPANY #1220 NORTHGATE ELECTRIC DOORS INC C&S BUILDERS SUPPLY #1006800 ATLAS DOOR CORP RCS SOUTHSIDE List of Nodes (Zipcodes with summary data) Break Flag Into Feasible Flags F2 Improve Set of Flags F3 Current Set of Flags Optimized Set of Flags Save Improved Flags F4 Orders Updated with Flag ID and Stop Sequence LOWES #683 ROSE CITY GARAGE DOORS INC Production-Transportation Scheduling 35 Production-Transportation Scheduling 36 Feasible Flag Formation Carrier Choice Orders to Flags F2 Initial Set of Flags Sort Flags by Azimuth F31 Current Set of Flags (Sorted) Let T = ( 0,1,..., t , K , N ) denote a tour of cities beginning with the depot. Let c( s, t ) denote the cost of a tour ( 0, s, s + 1,..., t ) . Let f ( t ) denote the minimum cost of serving cities ( t , K , N ) assuming that t is the first city to be visited after the depot and that the order of visiting cities on any subtour follows the order given by T . Dynamic Programming Recursion: f ( s) = min (c( s, t ) + f ( t + 1)), s ∈ T ; s≤t≤N t s t+1 Flag Improvement Cycle Next Flag to Consider Pick Next Flag F32 Current Azimuth Carrier Choice Orders to Flags Pick a Nearby Flag F33 Count Nearby Flag Combine the Two Flags F34 List of Nodes F1 (Zipcodes with summary data) Combined Flag f ( N + 1) = 0. Note: c(s,t) can include nested optimization of carrier choice and penalties for violating limits on vehicle loads, driver hours, etc. Break Combined Flag into Feasible Flags F35 New Flags Current Set of Flags Replace Original Two Flags if Improvement F36 Optimized Set of Flags F4 Production-Transportation Scheduling 37 Current Set of Flags Download Current Orders A1 Current Orders and Line Detail Scheduling Cycle A2 PTS Database Production Smoothing Index Sets S: the set of all schedule types; D: the set of all production days; F: the set of all selected flags; Parameters M: a large number; pfd : delay penalty for scheduling flag f on day d; afs : production (in linear feet) of schedule type s required by flag f; lsd : minimum desired production of schedule type s on day d; usd : maximum desired production of schedule type s on day d; Variables Xfd : = 1 if flag f is assigned to day d; 0 otherwise; Osd : excess production (in linear feet) of schedule type s on day d; Usd : under-production (in linear feet) of schedule type s on day d; User-Maintained Tables (Flag names, Carrier names, Pooling zipcodes) Build PTS Database Form Loads and Sequence Stops A3 Order and Line Detail Views, Selection Queries, Hilighting and Subselections Map View, Map Selection Operations Stop Sequence Optimization Flag Formation Optimization Unscheduled Flags Flags to Days Flags to Days Assign Flags to Production Days A4 Scheduled Flags Unscheduled Flag View, Scheduled Flag View, Drag and Drop Summary Totals View Export Revised Order File A5 Production Smoothing Optimization Revised Order File Upload Revised Orders A6 Production-Transportation Scheduling 40 Production Smoothing Choose ((X fd ), (Osd ), (U sd )) to minimize Z= d ∈D f ∈F LP Formulation Production Smoothing Optimization ∑∑ p fd fd X fd + ∑∑ M (Osd + U sd ) d ∈D s∈S subject to : d ∈D Flags to Days ∑X = 1, for all f ∈ F ; f l sd ≤ ∑a f ∈F X fd − Osd + U sd ≤ u sd , for all s ∈ S , d ∈ D; X fd ∈ {0,1}, for all f ∈ F , d ∈ D; Osd ≥ 0, for all s ∈ S , d ∈ D; and U sd ≥ 0, for all s ∈ S , d ∈ D. • Custom code for formulating LPs • Block form representation • Unlike AMPL (algebraic) • Each block defined using SQL queries 41 Production-Transportation Scheduling 42 Production-Transportation Scheduling Rolling Horizon Simulation • Month of April 1996 • Order availability and due dates deduced from actual shipments • Maximum 8-day order visibility • Due dates strictly enforced • Production targets mildly enforced Restrictions • • • • • • No backhaul opportunities Truck capacity 200 doors Max 5 drops for contract carrier Max 24 drops for fleet truck Max 2600 miles for contract carrier Max 2000 miles for fleet truck Production-Transportation Scheduling 43 Production-Transportation Scheduling 44 Simulation Results Cost Comparison of Historical vs. Optimized Flag Creation: April, 1996 Production Loading 250000 200000 FLEET CONTRACT FLEET CONTRACT Total Miles Miles Drops Drops Dollar Value Historical 282,046 269,005 2,146 625 $878,484 271,693 2,249 591 $797,591 Optimized 221,047 $1.38 $1.60 $9.95 $60.00 Cost Rate Per Cent Reduction 9.21% Linear Feet 150000 100000 Linear Feet Due Simulation Linear Feet Actual Linear Feet 50000 0 04/01/96 04/03/96 04/05/96 04/09/96 04/11/96 04/15/96 04/17/96 04/19/96 04/23/96 04/25/96 04/29/96 Date Production-Transportation Scheduling 45 Production-Transportation Scheduling 46 Overview • • • • • • Production Highlights Transportation Highlights Proposed System Optimization Decision Support System Design Principles Decision Support Carrier Choice Stop Sequence Orders to Flags Flags to Days Flags to Shifts Operation Sequence Routing Choice Execute Interactive Display Edit Interactive Execute Display Edit Production-Transportation Scheduling 47 Production-Transportation Scheduling 48 Indicators of Success • Schedulers loved the software • Days of work collapsed to hours • All relevant data at their fingertips • Software implemented at second plant • Software in daily use 1998-2005 • Company did not build finished goods warehouse Production-Transportation Scheduling 49 Production-Transportation Scheduling 50 Overview • • • • • • Production Highlights Transportation Highlights Proposed System Optimization Decision Support System Design Principles Design Principles for Decision Support Developers • It's the user's decision, not yours • Don't force a solution on the user • The user is not an analyst • Develop the interaction • Separate the model from the data • Prefer table lookups to hard-coded numbers • Prefer calculation queries to tables • Minimize the maintenance • Apply systems engineering • Capture the voice of the customer • Develop diagramming expertise • Divide and conquer Production-Transportation Scheduling 51 Production-Transportation Scheduling 52 Design Principles • Follow structured programming • • • • Software System Components PTS User Interface (Visual Basic) 3rd Party Components TransCAD Server (True DBGrid, Pinnacle Graph, Anibutton) Use short routines and long names Add error handling for all resources Don't hard-code path names: use .ini files Exploit component architectures • Buy most but write some Microsoft Jet Database Engine Microsoft Access SQL Queries Microsoft Access Microsoft Access Database Macros LPRun Optimization • Allow for customization • Queries, macros, reports • Get help (sub-contract or team) • Graphic art • Documentation • Technical programming PTS Mapping Macros (TransCAD UI Database) Geographic Files and Data Files PTS Databases XA Optimization Library Production-Transportation Scheduling 53 Production-Transportation Scheduling 54 Queries Database Macros Production-Transportation Scheduling 55 Production-Transportation Scheduling 56 TransCad Macros Programming a Simulation in Visual Basic Production-Transportation Scheduling 57 Production-Transportation Scheduling 58 Accessing a Database in VB Programming the User Interface Production-Transportation Scheduling 59 Production-Transportation Scheduling 60 Hire a Documenter The Production Transportation Scheduling System Carrier Choice PTS Stop Sequence Orders to Flags Flags to Days Flags to Shifts Operation Sequence Routing Choice Interactive Execute Display Edit PTS Sim Execute Interactive Display Edit Production-Transportation Scheduling 61 ...
View Full Document

This note was uploaded on 01/09/2009 for the course ORIE 312 taught by Professor D.ruppert,p.jacks during the Spring '08 term at Cornell University (Engineering School).

Ask a homework question - tutors are online