Optimization in Supply Chain Planning

Optimization in Supply Chain Planning Dr. Christopher Sürie Expert Consultant SCM Optimization

Agenda Introduction Hierarchical Planning Approach and Modeling Capability Optimizer Architecture and Optimization Strategies Customer Cases System Demo SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 2

Introduction: Supply Chain Management Supply Chain Management: Set of approaches utilized to integrate suppliers, manufactures, warehouses and stores so that merchandise is produced and distributed with the correct quantity to/from the correct locations at the correct time in order to minimize cost while satisfying service level requirements Prerequisite: Integrated Supply Chain Model Supplier Plant DC Customer SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 3

Introduction: Supply Chain Management Master Data Model Location (Plant, DC, Supplier,...) Lane Product Production Process Model (PPM) Resource Operational Data Demands Orders Capacity Profiles Resource Product Supplier Plant DC Customer SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 4

mysap SCM Solution Overview Measure Supply Chain Performance Management Supplier Supply Chain Exchange Network Partner Collaborate Supply Chain Collaboration Strategize Source Supply Chain Design Direct Procurement Make Plan Manufacturing Demand and Supply Planning Deliver Order Fulfillment Supply Chain Collaboration Collaborate Customer Supply Chain Exchange Network Partner Track Supply Chain Event Management SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 5

Supply Chain Planning Matrix Procurement Production Distribution Sales longterm Strategic network planning Master planning Demand planning Material requirements planning Production planning Distribution planning shortterm Detailed scheduling Transportation planning Available to promise (Stadtler/Kilger, 2005) SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 6

How to deal with planning complexity? Basic idea: Hierarchy of relaxations Relaxations are derived by Aggregation Time Periods Product Product groups (e.g. ignore country specific documentation in packaging a product) Resource Resource Families (e.g. summarize similar resources into one resource with cumulative capacity) Locations Regions (e.g. aggregate different locations into a transportation zone (postal code areas) Integration between different relaxations: Disaggregation SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 8

Hierarchical Planning Strategize Supply Chain Supply Design Chain Design Strategize Plan Plan Demand and Demand and Supply Chain Design Supply Planning Supply Planning Source Direct Procurement Make Manufacturing Deliver Order Fulfillment SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 9

Supply Network Planning (SNP) Strategize Plan Source Supply Chain Design Supply Network Planning (SNP) Combined production and distribution planning Mid-term Demand to long-term and planning horizon Quantity-based Supply Planning Aggregation Make Deliver Time Buckets, max. daily precision Direct Products, Order Manufacturing Resources Families Procurement Fulfillment SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 10

Supply Network Planning Procedures SNP Heuristics Material availability constraints Rule-based First feasible plan CTM (Capable To Match) Material availability and production capacity constraints Constraint-based propagation with backtracking search First feasible plan SNP Optimizer Material availability and all capacity constraints (MI)LP and others Cost-based optimization SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 11

SNP Optimizer Application Sourcing Product-Mix Which products and how much of them should be produced, transported, procured and stored? Technology-Mix Which recipes (PPMs) should be applied? Which transportation type should be used? Which resource should be used? Temporal: When should we produce, transport, procure and store? Spatial: Where should we produce, procure and store? Wherefrom and whereto should we transport? Finite Planning Lot-Sizing Multi-Level-Capacitated Lot Sizing (MLCLSP) Campaign Planning Inventory Control Target Days of Supply Shelf Life SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 12

Supply Network Optimization: Model Building Transport Capacity Transport Discrete Lots Minimal Lots Piecewise linear Costs Produce Discrete Lots Minimal Lots Fixed Resource Consumption Piecewise linear Costs Production Capacity PPM Storage Capacity Products Safety Stock Handling-In Capacity Handling-Out Capacity Store with Shelf life Procure Piecewise linear Costs Satisfy Demand with Demand Classes Delay Costs Non-Delivery Costs SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 13

SNP Optimization Run Demand Forecast SNP Optimization Resource Selection SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 14

Supply Network Optimization: Lot-Sizing Multi-Level Capacitated Lot Sizing Problem (MLCLSP) Setup cost and/or consumption in each bucket Good results Setup cost small compared to storage cost ( Small lots) Setup consumption << bucket capacity Bad results Setup cost large compared to storage cost (large lots) Setup consumption big compared to bucket capacity SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 15

Supply Network Optimization: Lot-Sizing Proportional Lot Sizing Problem (PLSP) Setup cost and/or setup consumption only if different PPM starts At most one startup per bucket Constraints on cross-period lots (= campaign quantity) Minimal campaign quantity Campaign quantity integer multiple of batch size SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 16

Manufacturing (PP/DS) Strategize Supply Chain Design Plan Source Manufacturing (PP/DS) Direct Procurement Demand and Supply Planning Combined material and capacity planning Make Manufacturing Deliver Short-term to mid-term planning horizon Order based Time continuous Order Fulfillment SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 17

PP/DS Planning Procedures PP/DS Heuristics Material availability, single-level finite Priority-based planning First partial plan CTM (Capable To Match) Material availability and production capacity constraints Constraint-based propagation with backtracking search First feasible plan PP/DS Optimizer All constraints Genetic Algorithm and Constraint Programming Cost-based Optimization SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 18

PP/DS Optimizer Application Feasible compact schedule Delay Reduction Makespan Minimization Setup Minimization Time Cost Resource Selection SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 19

PP/DS Optimizer: Model Overview Setup sequence dependent setup costs Time Windows earliest starting time due date, deadline delay costs Distances (with calendars) minimal maximal Resource Selection Alternative resources Resource costs Unary Resources Storage resources Product Flow discrete continuous Multi-Cap Resources SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 20

PP/DS Optimization Run SNP PP run DS optimization SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 21

Integrated hierarchical planning SNP Planning only in SNP horizon Release SNP Orders only PP/DS horizon Respect PP/DS orders as fixed capacity reduction material flow Respect PP/DS setup state PP/DS Respect pegged SNP Orders as due dates No capacity reduction But material flow No restrictions for scheduling PP/DS orders SNP PP/DS PP/DS Horizon SNP Horizon SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 22

VSR: Model Building Classical vehicle routing problems: CVRP, CVRPTW m vehicles 1 2 3 6 demand b j 4 5 8 9 time window [l j,r j ] capacity c 7 supply a time window [l,r] SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 23

VSR: Model Building (ct d) Extensions towards APO VSR (1): Order-based model Source location and destination location per order (Pickup and delivery problem) 1 5 O4 6 O1 2 O2 3 O3 4 Quantity regarding loading dimensions (tons, m 3,...) Material type (chemicals, food,...) Service times for loading and unloading (depends on vehicle) SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 24

VSR: Model Building (ct d) Extensions towards APO VSR (2): Cost for not delivering an order Soft/hard time constraints per order: Earliest date for pickup Due date for pickup Earliest date for delivery Due date for delivery cost Early Pickup [ ] [ ] Late Pickup Early Delivery Late Delivery time SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 25

VSR: Model Building (ct d) Extensions towards APO VSR (3): Per vehicle: Travel characteristics (time, distance per lane) Start location and end location Constraints Capacity per loading dimension (tons, m 3,...) Limit for time, distance, number of stops Break calendar Costs: Fixed cost Traveled time Traveled distance Number of stops Distance x Load (e.g. miles x tons) Vehicle type = vehicles with identical travel & cost characteristics SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 26

VSR: Model Building (ct d) Extensions towards APO VSR (4): Per location: Deliveries require inbound resource Opening times Capacities Pickups require outbound resource Opening times Capacities SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 27

VSR: Model Building (ct d) Extensions towards APO VSR (5): Incompatibility constraints: Between material types Between vehicle types and material types Between vehicle types and locations Schedule vehicles (e.g. trains, ships) Route and schedule is fixed a priori Hubs Indirect shipment through hub(s) versus direct shipment Maximum waiting time at hub 1 H2 H1 2 SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 28

Challenge: Generic Optimizer Generic and SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 30 Best of Breed planning level vertical industries run time requirement model complexity (size, constraints, objectives) Generic Model (-> planning level) aggregated planning (LP / MILP) detailed planning (scheduling) Customization (-> vertical industries) specialization the generic model to customer problem scripting the strategies (decomposition, goal programming) Scalability (-> run time) greedy versus complex optimizations strategies parallelization Open Architecture internal: adding new special optimizer (software evolution) external: integration of optimizer packages

SNP Optimizer Architecture GUI Reporting Checking Control LiveCache/DB Model Generator Core-Model Time- Decomposition Product- Decomposition Priority- Decomposition Resource- Decomposition Meta-Heuristics SNP LP/MILP SNP Rule based Deployment LP/MILP Basic-Optimizers SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 31

Scheduling Optimizer Architecture LiveCache GUI Model Generator Time Decomposition Reporting Bottleneck Checking Core Model Multi Agent Control Meta-Heuristics Constraint Programming Genetic Algorithm Sequence Optimizer Campaign Optimizer Basic Optimizer SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 32

Mastering the algorithmic complexity: Decomposition Global versus local optimality -> SNP + DS Local optimality depends on neighborhood High solution quality by local optimization Local Optimization = Decomposition Decomposition strategies SNP: time, resource, product, procurement DS: time, resource (Parallelization by Agents ) SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 33

SNP Product Decomposition SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 34

SNP Time Decomposition 1 2 3 4 5 6 solved solved merge 1 2 3-6 Time- Decomposition extract solve SNP LP/MILP store SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 35

DS Time Decomposition - Local Improvement Resources Current window Time Gliding window script 1. Optimize only in current window 2. Move window by a time delta 3. Go to first step SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 36

DS Metaheuristics - Bottleneck Resources Bottleneck Time Bottleneck Script 1. Determine bottleneck 2. Schedule bottleneck resources only 3. Fix sequence on bottleneck resource 4. Schedule all resources SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 37

VSR: The Optimizer Generic Optimizer Preprocessing Which orders cannot be delivered at all? Which order can be processed by which vehicle? Postprocessing Shift travel activities forward or backward S-1 P1 1-2 D1 (forward) S-1 P1 1-2 D1 (backward) SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 38

VSR: The Optimizer (ct d) Evolutionary local search (ELS) with small population (3) Uses GENEAL (GENeral Evolutionary Algorithm Library) Direct solution representation Assignment of orders to vehicles Routing of activities on vehicles Scheduling of activities on vehicles Each atomic move has three phases: 1. Change assignment 2. Change routing 3. Change scheduling 19 atomic moves, classified into Assignment moves Routing moves Scheduling moves SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 39

Challenges in modeling real-world problems Business acceptability of computed solutions Solution quality gap (fixed run-time) Model too simplified Optimal model detail Model too detailed Model detail Generic nature of the SNP optimization model restricts exploitation of specific problem structure SAP AG 2005, Optimization in SC Planning, Dr. Christopher Sürie, 41