Warehousing Systems Design

Transcription

1 Warehousing Systems Design Marc Goetschalckx, Doug Bodner, T. Govindara, Leon McGinnis, Gunter Sharp, Kai Huang Industrial and Systems Engineering Georgia Institute of Technology Atlanta, GA (404) Credits Faculty L. McGinnis, D. Bodner, T. Govindara, M. Goetschalckx, G. Sharp Graduate students K. Huang Sponsors Keck foundation, SF, Ford Motor, UPS WWL, rights reserved.

2 Distribution Warehouse: Functional Areas Illustration Changing Warehousing Requirements Faster turns, shorter life cycles Warehouse operations Warehouse design Proliferation of information technology WMS, tracking Design-as-usual rights reserved. 2

3 Warehousing Design Obective The goal of warehouse design is to minimize the discounted present value of the costs of establishing and operating the warehouse over some horizon specified by the decision-maker, subect to a number of resource and performance constraints. State-of-the-art in Warehousing Design: Review and Design Reviews, bibliography, and design Van den Berg & Zim (999) Ackerman (999) Rouwenhorst at al. (2000) Miller (200), chapter 5 rights reserved. 3

4 Warehousing Types Distribution Consolidation Production WIP and finished goods storage Long-term storage Archival, waste and emergency storage Contract TPL Warehousing Levels of Automation Manual Picker to product, shelves Mechanized Picker to product, person-aboard Automated Product to picker, carousel Automatic Product to shipping, A-frame rights reserved. 4

5 Warehousing Design: Current Characteristics Overwhelming complexity and variety o monolithic model Hierarchical, iterative models Approximate analytical models Dramatically limit number of alternatives Final choice based on detailed simulation Warehousing Design Hierarchy: Strategic Decisions Strategic: configure the infrastructure Mission for the warehouses (if present) Owned versus contracted (TPL) umber, location, and function of the warehouses Technology and size of the warehouses rights reserved. 5

6 Warehousing Design Hierarchy: Tactical Decisions Tactical: utilize the infrastructure through policies Capacity balancing in the network Customer and product allocation Maor operational policies Warehousing Design Hierarchy: Operational Decisions Operational: schedule activities Operating procedures Short term scheduling Design feedback loops rights reserved. 6

7 State-of-the-art in Warehousing Design: Analysis Analysis of isolated components umerous, focused, uneven Detailed simulation and animation Clearly suboptimal o impact on the practice of warehousing design Review of Published Research Progress in Material Handling Research (998) State-of-the-art in Warehousing Design Continued o structured, integrated design methodology o model for warehousing operations o explicit design theory Consulting-university cooperatives for DSS and education Twente, Dortmund, Darmstadt, Georgia Tech, Rensselaer rights reserved. 7

8 Warehousing Design Methodology eeds Integration of isolated research and methods Rich empirical data sets Rigorous mathematical models Synthesis and design tools Warehousing Design and Operations Education Status Mirrors state-of-the-art Focused, uneven, analysis High demand Bachelors, graduate, and professional Few educational materials MHI and CICMHE (CD on MH principles) rights reserved. 8

9 Warehousing Design Development at Georgia Tech Functional flow network Obect oriented data base Mathematical design models Visualization Warehousing Functions Receiving Storage - holding Order picking - retrieval Consolidation - sorting Shipping rights reserved. 9

10 Functional Flow etwork for Basic Warehousing Functions Inspection Receiving Order picking Bulk storage Shipping Carton storage and picking Order sort/ accumulate Material flow Function-to-space mapping Functional Flow etwork Characteristics Every warehouse has a different functional flow network Different products can follow different and multiple paths Functions are mapped to areas, areas are mapped to material handling and storage equipment rights reserved. 0

11 Obect Classes in Warehouse Design Proect Obect (Design Control) Warehousing Obects (Common Building Blocks) Functional Flow Obects (Requirements and Products ) Operation Obects (Warehouse Configuration and State) Obect Classes Design Obect Design proect status Warehousing Obects Equipment and policies and protocols Functional Flow Obects Products and flows requirements Operations Obects Design specification rights reserved.

12 Warehousing Obects Structure (Equipment and Personnel) Block Stacking Example: Software Program Interactions 2. Block Stacking Algorithm. User Inputs the Region Information (VISIO) Unix/PostGres 3. Display Results (VISIO + VRML) rights reserved. 2

13 Department Block Layout in Visio Block Stacking Layout in VRML rights reserved. 3

14 Warehouse VRML Visualization: Side View with Storage Systems Warehouse VRML Visualization: Top View with Storage Systems rights reserved. 4

15 Warehouse VRML Visualization: Interior View Design Optimization Model Global Decisions Receiving Decisions Bulk Store Decisions Shipping Decisions Total Cost Function Global Constraints: Space, Budget, etc Receiving Bulk Store Constraints Shipping rights reserved. 5

16 Optimization Model Characteristics Optimization decomposed by functional areas Joint space (area), manpower, budget constraints modeled as resources Economic analysis oint obective function Optimization Model Time Units Two fundamental time units Epoch (year) Time period Discount factor and frequency Year e = Year 2 e = 2 Mon t= Tue t=2 Wed t=3 Thu t=4 Fri t=5 Sat t=6 Sun t=7 Mon t=8 Tue t=9 Wed t=0 Thu t= Fri t=2 Sat t=3 Sun t=4 Year e = Year 2 e = 2 Winter 200 t= Spring 200 t=2 Sum. 200 t=3 Fall 200 t=4 Winter 2002 t=5 Spring 2002 t=6 Sum t=7 Fall 2002 t=8 rights reserved. 6

17 Functional Areas Set of candidate technologies L Binary selection variables y Continuous flow variables v Investment cost, fixed cost, variable cost Individual commodity capacity constraints Functional Areas Equations = l L = l L t= = l L t= TechInvestCost T T y cdf freq TechFixedCost y l t t lt l cdf freq TechFlowCost v t t lpt lpt v TechFlowCap y lpt lpt l rights reserved. 7

18 Resource Equations (Flow) R T r= = l L t= = l L p= P cdf freq ResProdCost v t t rlpt lpt ResProdReq v ResCap rlpt lpt rt Throughput and Conservation of Flow Throughput requirements by commodity and type Three types of conservation of flow. Simple 2. Simple + inventory 3. Simple + inventory + transformation rights reserved. 8

19 Flow Requirements and Type Simple Conservation l L v FlowReq, p, t lpt x = v, p, t ipt i= l L x = v, p, t kpt k= l L lpt lpt pt Type 2 Inventory Conservation of Flow ipt i= l L + lp( t ) = + kpt k= l L lpt lpt lpt lpt l L l L l L l L x u u w v w x = = v lpt rights reserved. 9

20 Type 3 Transformation Conservation of Flow ipt i= l L + lp( t ) = + kpt k= l L p lpt lpt lpt lpt l L l L l L l L x q COMP u u w v w x = = v + lpt ProdTransform qp lqpt lpqt q ASSY p Inventory Equations = l L t= r= = l L t= = l L p= T t t lpt lpt lpt lpt l R T P cdf freq TechStoreCost w w TechStoreCap y P T p= = l L t= cdf freq ResProdCost w t t rlpt lpt ResProdReq w ResCap rlpt lpt rt cdf freq h ProdValue w t t t pt lpt rights reserved. 20

21 Transformation Equations TransformCap y lpqt lpqt l P P T p= q= = l L t= cdf freq TransformCost t t lpqt lpqt Iterative Warehouse Design Algorithm Solve capacitated warehouse flow model (CMF) MIP Determines flows, technologies and areas Solve conceptual block layout (WBL) Block layout heuristics or MIP Determines location, transportation costs Iterate rights reserved. 2

22 Conceptual Layout Illustration Transportation Equations P T p= i= = t= R T r= i= = t= P i= = p= cdf freq TransportCost x t t ipt ipt cdf freq ResProdCost x t t ript ipt ResProdReq x ResCap ript ipt rt rights reserved. 22

23 Future Research Draft paper and presentation on the web: Computational test cases Please lend us your data!! Thank You Can I Answer Any Questions? rights reserved. 23