Post on 20-Aug-2015
CastilloCastillo
ChuaChua
IbuyanIbuyan
Interesting Figures…Interesting Figures…
$400 Billion
$178 Billion per year
$40 Billion
80%$6 - $8 Billion per year
Challenges of Spare Challenges of Spare PartsParts
• High Demand Uncertainty
• Increase in Prices for Individual Parts
• Higher Service Requirements
• Financially Remarkable Stock-out Effects
• Large Variety
• Slow Usage
• High Criticality
• Obsolete Factor
Challenges of Short Challenges of Short Lifecycle ProductsLifecycle Products
• Inventory
• Procurement
• Capacity
• Forecast
• Disruptions
• Competition
• Long lead-times of components and spare parts
• Uncertainty in the prices of components and spare parts
• Unpredictable market response
IBM Global Chief Supply Chain Officer Study 2010
What has been done What has been done so far? so far?
Review of 200 Articles Revealed Main Themes of Researches
1.Risk minimization models that do not incorporate multi-echelon inventory equations.
2. Studies on spare parts multi-echelon models that do not attempt to understand system behavior when subjected to uncertainty.
Main Contributions to Main Contributions to Spare Parts Studies Spare Parts Studies
1. Easily Extensible Spare Parts Multi-echelon Multi-Indenture Model that Incorporate Risks
2. Alternative Solution Method: Excel Based Genetic Algorithm
3. Results of Analysis of Spare Parts Systems when Subjected to Uncertainty and Risks
Sherbrooke (1968; 2004); Kutanoglu et al (2005, 2007)
Sherbrooke (1968; 2004); Kutanoglu et al (2005, 2007)
Excel-based Genetic Algorithm (VBA)
Excel-based Genetic Algorithm (VBA)
3 RSM Experiments (2 Robust Design Experiments)• Demand Variability• Cost Variability
3 RSM Experiments (2 Robust Design Experiments)• Demand Variability• Cost Variability
200 Articles Reviewed200 Articles Reviewed
Risks Scenario Analysis Risks Scenario Analysis
• General Guidelines for Spare Parts Supply Chain Design
• Optimal Risk Mitigation Strategies
• General Guidelines for Spare Parts Supply Chain Design
• Optimal Risk Mitigation Strategies
Model Overview
System Considered
Inventory Control Point for Service Parts
Flow of Usable Spare Parts
Flow of Unrepaired Spare Parts
Regular Transport
Emergency Transport
Plant n
Retailer 1
Retailer 1
Retailer 2
Retailer 2
Retailer m
Retailer m
DC 1
DC 2
Plant 1
Retailer 3
Retailer 3
DC i
::
:
System DiagramSystem Diagram
Level 0 Part
Level 1 Part
Level 2 Part
Multi – Indenture Multi – Indenture Products SystemProducts System
Retailer DC Plant
Level 0 Part
Level 1 Part
Level 2 Part
Level 0 Part
Level 1 Part
Level 2 Part
Level 0 Part
Level 1 Part
Level 2 Part
Multi – Echelon Inventory Multi – Echelon Inventory Equation DevelopmentEquation Development
Solution Methodology
Excel-based Genetic Algorithm
Excel Template
Coding the Mathematical Model
VBA for Genetic Algorithm
Genetic Algorithm Parents Selection
Crossover Technique
Mutation Mechanism
Analysing Genetic Algorithm Behavior Over Time
Results
Outputs of Research
1. General Guidelines for Incorporating Risk to Spare Parts Supply Chain Design
2. Table of Risk Mitigation Strategies to Hedge against Different Types of Risks
General Guidelines for Robust General Guidelines for Robust Service Parts Supply ChainsService Parts Supply Chains
1. Always choose to place inventory at downstream elements first.
2. Lowering delays anywhere inside the system lead to more desirable supply chains in terms of cost.
3. While considering Rule 1, it is also always better to locate most of the inventory at facilities where majority of the repair operations are done.
4. Stock low criticality spare parts in the most upstream element of the supply chain.
5. Optimize the allocation of high criticality parts across all elements of the supply chain.
When Facing Demand Variability – Preventing “Bullwhip Effect” in Spare Parts
Systems
1. Lower delay times (ex. repair times) become more crucial in the face of demand variability.
2. Reduce delays at ALL facilities of the Spare Parts Supply Chain, even the facilities that receive little demand.
3. Repair should be done as much as possible at downstream elements to minimize total cost and its variability.
General Guidelines for Robust General Guidelines for Robust Service Parts Supply ChainsService Parts Supply Chains
When Facing Cost Variability
1. General Rules in Optimizing Service Parts Supply Chain Apply.
2. Unlike in the situation of demand variability, being robust to cost variability only requires the supply chain to minimize delays at FACILITIES WITH HIGH DEMAND.
General Guidelines for Robust General Guidelines for Robust Service Parts Supply ChainsService Parts Supply Chains
Risk Scenario 1st best Strategy 2nd best Strategy 3rd best strategy
High Demand Variability
Increased Responsiveness
Increased Inventory in all Sites
Increased Inventory at Plant and Retailer
Inventory Limit Increased Responsiveness
Increased Inventory at the DC
Increased Inventory at the Plant or at both Plant and DC
Extreme Demand Values
Increased Responsiveness
Increased Inventory in all Sites
Increased Inventory at DC and Plant
Inventory Cost Increased Responsiveness
Increased Inventory in all Sites
Increased Inventory at Plant and Retailer
Facility Cost Increased Responsiveness
Increased Inventory in all Sites
Increased Inventory at Plant and Retailer
Emergency Shipment Cost
Increased Responsiveness
Increased Inventory in all Sites
Increased Inventory at Plant and Retailer
Results of Risk RunsResults of Risk Runs
Total Average Cost Total Average Cost Across ScenariosAcross Scenarios
• Most effective Risk Mitigation strategy.
• Short lifecycle products are best served by a responsive supply chain (Cohen et al, 2006).
• Quick response to demand lessens penalty costs and fulfills a high service level requirement.
Increased Increased ResponsivenessResponsiveness
RecommendationsRecommendations
• Design of a C++ program that can execute the genetic algorithm faster.
• Adding more scenarios and conducting a full stochastic programming analysis.
• Design of alternative solution methodology to solve the mathematical model.
• Designing mathematical functions that can better approximate the pipeline inventory.