Presentation is loading. Please wait.

Presentation is loading. Please wait.

Evaluating and Optimizing the Performance of Complex Multi-stage Supply Chains Under Disruptions Sanjay Kumar University of Texas at Dallas Kathryn E Stecke.

Similar presentations


Presentation on theme: "Evaluating and Optimizing the Performance of Complex Multi-stage Supply Chains Under Disruptions Sanjay Kumar University of Texas at Dallas Kathryn E Stecke."— Presentation transcript:

1 Evaluating and Optimizing the Performance of Complex Multi-stage Supply Chains Under Disruptions Sanjay Kumar University of Texas at Dallas Kathryn E Stecke University of Texas at Dallas Thomas G Schmitt University of Washington Seattle Fred Glover University of Colorado at Boulder

2 A Simple Multi-Stage Supply Chain SuppliersManufacturersDistributorsRetailersCustomers Transporters

3 Demand Supply Chain Management Certain events can disturb the balance of demand and supply. Supply

4 A Simple Multi-Stage Supply Chain SuppliersManufacturersDistributorsRetailersCustomers Transporters Natural catastrophes Accidents Terrorist attacks Modern supply chain design

5 Scope of this Research Develop optimization tools for making cost-effective decisions under disruptions. Study the effectiveness and rationale of popular disruptions response methods used in supply chains. Explore the vulnerabilities and understand the (long-term) effects of disruptions at various stages of a supply chain. Supply Chain Risk Management Ordering decisions under disruptions

6 Outline of the Presentation Problem background and motivations The model Literature Solution methodologies Results Conclusions and contributions

7 Background and Motivations

8 Recent Supply Chain Disruptions 9/11 –Economic losses to New York city in the month following the attacks: $1.5 billion –Jobs lost in NY: 200,000 –Estimated total jobs lost in the country: 1.5 million Hurricane Katrina –Economic losses to insurance industry far exceeded the losses because of hurricane Andrew, 9/11, and Northridge California earthquake combined together. The 2000 fuel crisis in UK –Resulted in disruptions far more severe than 9/11. Various types of disruptions affect supply chains. For many industries 9/11 was not the most disruptive event.

9 Disruptions Response Decisions made during disruptions are often based on short-term goals, or lack of foresight. In many cases losses occur because of “poor” or “wrong” response –9/11 and Homeland Security Advisory System –Kobe earthquake Does company level decisions made during disruptions also negatively affect the supply chain performance? –Ordering and transportation

10 Disruptions Response Homeland Security Department: –Sandia National Labs started developing models to understand the economic consequences of disruptions in critical infrastructure. –The aim was to predict and mitigate the economic effects of disruptions in Manufacturing facilities Transportation Electric power Telecommunications

11 Manufacturing/Transportation: Questions to Address What kind of disruptions affect manufacturing/ transportation? –Length –How often How does present supply chains cope with them? Can we help companies make better ordering decisions during disruptions and even otherwise? Does company level decisions made during disruptions also negatively affect the supply chain performance? –Ordering and transportation Answers to these questions could vary between industries.

12 Our focus Electronics companies

13 Why Electronics Manufacturing Supply Chains? Electronics are widespread in the functioning of our society. –Since WWII, electronics products have accounted for over 30% of US GDP. Electronics assembly is very susceptible to disruptions. –Typical electronics products can have 70-700 components Electronics supply chains involve global, multinational interests that broaden the exposure to disruptions. –Over 80% of electronics components are internationally sourced. Modern electronics products have very short life cycles. –Less than 4 months for DVD players and Digital Camcorders

14 Key Characteristics of an Electronics Supply Chain: Three Case Studies (from a sample of 14,000 electronics firms) Design of supply chain –Assembly is an integral part. –Often global. Response –Each company expedite orders to overcome shortages. The final customer demand follows AR(1) process. –The demand across periods are correlated. Supply chain well coordinated –Shortages become lost sales only at the retailer.

15 Level 1 Supply Chain Level 4 Level 3 Level 2 LT: 10 ELT: 6 LT: 30 ELT: 15 LT: 25 days ELT: 10 days LT: 35 ELT: 30 LT: 42 ELT: 40 Assembly (Finished Product) Stage 4A LT: 45 ELT: 40 Stage 3A LT: 30 ELT: 28 Assembly is an integral part of electronics supply chain. Both facility and transportation disruptions are critical. Each stage expedites orders to overcome shortages.The final customer demand is AR(1). The supply chain is well coordinated. Demand is lost only at the retailer. Stage 3B Stage 4B SuppliersManufacturersDistributors Retailer Assembler

16 Problem Statement In a multi-stage model supply chain, determine the cost effective order-up-to levels for each stage considering the costs of –Backorders –Lost sales –Inventory carrying –Expediting

17 Literature Supply chain security: CSI, increased tracking and visibility, product and process standardization (Sheffi, 2003). Inventory policies –Single stage –Stationary assumptions –Nonstationarity is induced by expediting and disruptions Little research to find policies for multi stage and non-stationary supply chains considering bullwhip. –Chen et al. (2001), Lee et al. (1997), and Kahn (1987) deal with the existence and quantification of bullwhip. Almost all research articles consider an i.i.d demand. –The “best” demand function is correlated across periods.

18 Objective Function A weighted function of the costs of expediting, backorder, lost sales, and inventory holding is minimized. s.t. Inventory flow constraints are satisfied (next slide) Decision variables: Order quantities at each of the six stages of the supply chain. Holding cost Backorder cost Lost sales cost Expediting cost

19 Flow Constraints (for Stage i) Inventory and quantity on order Previous Period Inventory Regular Shipment Expedited Shipment Inventory

20 Flow Constraints (for Stage i) Shipment to Stage i-1 Shipments are minimum of available inventory and the order quantity +backlog Additional constraint for assembly stage

21 Flow Constraints (for Stage i) Regular Shipment to stage i-1: Expedited Shipment to stage i-1: If inventory is positive, regular orders are placed Negative inventory (shortages) results in expedited orders

22 Flow Constraints (for Stage i) Shortages: Backorders and lost sales: Effective order- inventory All shortages backordered A fraction is backordered, rest is lost

23 Nature of the Cost Function

24 Solution Strategies The objective function is non-convex in the order quantities. Certain deterministic cases are NP complete. Solution methods –Fibonacci Results in local optimal solutions –Genetic algorithms Significantly longer run time –Tabu search

25 GA (139.7 min) Fibonacci (2.8 min) Tabu (28.5 min) Cost Comparison of the Solution Methods

26 No-expediting 4.7% No Expediting Cost Expediting vs. no Expediting

27 Effect of Assembly Assembly stage reduces the order amplification effect –The reduction is prominent in the higher stages. –The bullwhip-reducing effect of assembly increases with increase in number of components assembled. This provides an explanation for counter-intuitive results of Cachon et al. (2006). 2 components assembly No assembly

28 Effects of Disruption 15 days disruption at retailer Magnitude of losses 15 days disruption at manufacturer Magnitude of losses Disruption

29 Conclusions and Summary –We developed and implemented a search-based optimization methodology and effectively used it to find order-up-to quantities in a multi stage supply chain. First such method with the potential to help supply chains make ordering decisions considering -Nonstationarity -Expediting –Tabu Search First such application for Tabu search. Developed and adapted Tabu search to be effectively used for this problem. Genetic search is shown to be inferior.

30 Conclusions and Summary –Bullwhip Assembly stage filters the demand thus reducing bullwhip. We provided a possible explanation to Cachon et al.’s findings. –Expediting Widely prevalent expediting practice may hurt supply chain performance. Expediting may also result in longer recovery times.


Download ppt "Evaluating and Optimizing the Performance of Complex Multi-stage Supply Chains Under Disruptions Sanjay Kumar University of Texas at Dallas Kathryn E Stecke."

Similar presentations


Ads by Google