Post on 17-Dec-2015
Inventory routing problems- Combined inventory management and routing
Henrik AnderssonArild Hoff
Marielle ChristiansenGeir Hasle
Arne Løkketangen
DOMinant workshop - Molde - September 20-22 2009
Outline
• Background• Industrial aspects• Literature review and classification• Industrial aspects in the literature• Trends and future directions
Background
• DOMinant is a collaboration between Molde University College, Norwegian University of Science and Technology and SINTEF
• Discrete optimization models in maritime and road-based transportation– Fleet size and mix vehicle routing problem– Inventory routing problem
• First task : Get familiar with the research and write a survey on each problem
DOMinant workshop - Molde - September 20-22 2009
DOMinant workshop - Molde - September 20-22 2009
Surveys
• Focused surveys with a broader purpose– Overview of the diverse landscape of combined inventory
management and routing – Relation and interconnection between science and practice– Practitioners can see the benefit of using advanced decision support
systems in complex situations– Inspire researchers to dig deeper into the field and find new and
exciting challenges– Discuss both land-based and maritime applications
• Mainly based on the research literature – Own experience– No field study or industry involved in the writing process
DOMinant workshop - Molde - September 20-22 2009
Combining operations I
• Combining operations within the supply chain– Increased efficiency– Economical benefits– Flexibility in services– Improved robustness
• Which operations to integrate?• Who should plan the integrated operations?• For which products is it interesting to integrate operations?• When should operations be integrated?
DOMinant workshop - Molde - September 20-22 2009
Combining operations II
• Which operations to integrate?– Inventory ≈ A buffer between processes to even out variations and
handle uncertainty– Managing the conflicting goal between supply and demand
• Procurement – Inventory – Production (intra)• Production – Inventory – Sales (intra)• Inventory – Transportation – Inventory (intra and inter)
Supply
Demand
DOMinant workshop - Molde - September 20-22 2009
Combining operations II
• Which operations to integrate?– Inventory ≈ A buffer between processes to even out variations and
handle uncertainty– Managing the conflicting goal between supply and demand
• Procurement – Inventory – Production (intra)• Production – Inventory – Sales (intra)• Inventory – Transportation – Inventory (intra and inter)
Supply
DemandInventory
DOMinant workshop - Molde - September 20-22 2009
Motivation
Producer
Supplier Customer
Customer
Customer
Customer
Producer
Producer
Supplier
DOMinant workshop - Molde - September 20-22 2009
Motivation
Producer
Supplier Customer
Customer
Customer
Customer
Producer
Producer
Supplier
Control Execute
DOMinant workshop - Molde - September 20-22 2009
Motivation
• Economical benefits• Flexibility in services• Improved robustness
Producer
Supplier Customer
Customer
Customer
Customer
Producer
Producer
Supplier
ControlExecute
Combining operations III
• Who should plan the integrated operation?– The producer– A transport provider– A third actor
DOMinant workshop - Molde - September 20-22 2009
DOMinant workshop - Molde - September 20-22 2009
Combining operations IV
• For which products is it interesting to integrate operations?– High value products– Perishable products– High impact products
DOMinant workshop - Molde - September 20-22 2009
Combining operations V
• When should operations be integrated?– Autonomous actors in the chain• Economically beneficial at the system level• Possible to split the benefits in an agreeable way• Long-term relationship• Willingness to share information and data
DOMinant workshop - Molde - September 20-22 2009
Current industrial practice
• Inventory management and routing are treated as separate operations– manual planning is still used at many companies
• There are very few, if any, commercial optimization-based systems in current use
• Several integrated systems are under development for companies in the maritime sector– The liquefied natural gas industry mainly due to an
extraordinary strong growth in this high value business
DOMinant workshop - Molde - September 20-22 2009
Research perspective
• Looking at the coordination of inventory management and routing, it is clear that the combined problem is a long-term, dynamic problem that is inherently stochastic
• Since long-term, dynamic and stochastic problems are extremely difficult to solve, the approaches found in the literature have simplified the problem in one way or another
DOMinant workshop - Molde - September 20-22 2009
ClassificationCharacteristic Alternatives
Time Instant Finite Infinite
Demand Deterministic Stochastic
Topology One-to-one One-to-many Many-to-many
Routing Direct Multiple Continuous
Inventory Fixed Stock-out Lost sale Back-order
Fleet Homogeneous Heterogeneous
Fleet size Single Multiple Unconstrained
Mode Road-based Maritime Generic
Finite
• More than one visit at a customer may be needed
• Common characteristics– Deterministic demands– Fixed inventory restrictions
• Road-based, maritime and generic– Many-to-many topology without depot– Few deliveries– Heterogeneous fleet
DOMinant workshop - Molde - September 20-22 2009
DOMinant workshop - Molde - September 20-22 2009
Finite II
• Very few exact approaches– Christiansen and Nygreen, 2005 • Robust plans, penalize closeness to inventory limits• Path-based formulation• Branch-and-price
– Archetti et al., 2007• Order-up-to-level policy, one vehicle• Arc-flow formulation• Branch-and-cut
DOMinant workshop - Molde - September 20-22 2009
Finite III
• MP-based and neighborhood-based heuristics– Arc-flow
• Miller, 1987 ; Al-Khayyal and Hwang, 2007 ; Savelsberg and Song, 2008 ; Yugang, 2008
– Path-flow• Bell et al., 1983 ; Persson and Göthe-Lundgren, 2005
– Neigborhood-based• Rusdiansyah and Tsao, 2005 (Tabu search) ; Abdelmaguid
and Dessouky, 2006 (GA) ; Alegre et al., 2007 (Scatter search) ; Dauzère-Pérès, 2007 (Memetic) ; Savelsbergh and Song, 2007 (GRASP) ; Hemmelmayr et al., 2008 (VNS)
DOMinant workshop - Molde - September 20-22 2009
Finite IV
• Heuristics : Main approaches– Allocate first – route second• Carter et al., 1996 ; Campbell and Savelsbergh, 2004
– Periodic and delivery patterns• Gaur and Fisher, 2004 ; Rusdiansyah and Tsao, 2005 ;
Alegre et al., 2007
– Imposing different inventory policies• Kim and Kim, 2000 ; Bertazzi et al., 2002
DOMinant workshop - Molde - September 20-22 2009
Industrial aspects I
• Products having a relatively high consumption rate.– Gas, Bell et al., 1983, Golden et al., 1984, Campbell and
Savelsbergh, 2004 ; chemicals, Miller, 1987 ; automobile components, Blumenfeld et al., 1987, Alegre et al., 2007 ; ammonia, Christiansen, 1999 ; groceries, Gaur and Fisher, 2004 ; bitumen, Persson and Göthe-Lundgren, 2005 ; frozen products, Custódio and Oliveira, 2006 ; calcium carbonate slurry, Dauzère-Pérès et al., 2007 ; petrochemical products, Al-Khayyal and Hwang 2007 ; blood, Hemmelmayr et al., 2008
DOMinant workshop - Molde - September 20-22 2009
Industrial aspects II
• The uniqueness of the industrial cases is important– Factors not included in already existing models
• Can either complicate or sometimes simplify the problem
• Knowledge about the number of customers that are normally visited on a route to simplify the generation of routes
• Using the consumption/production rates and inventory bounds to explicitly model each possible port call and to derive time windows for each visit
• Characteristics from the current manual planning
DOMinant workshop - Molde - September 20-22 2009
Industrial aspects
• Most research where industrial problems are focused holds elements of heuristics– Exact methods terminated prematurely– Neighborhood-based heuristics and metaheuristics
• The diversity of methods used shows many paths for designing algorithms for industrial cases
• Knowledge about the case, and the delicate use of this knowledge when formulating and solving the problem shines through in most applications
DOMinant workshop - Molde - September 20-22 2009
Status summary
• Gain in interest due to increased focus on supply chain management
• Many different problems, no clear definition• A gap between research and academia• Little work on exact methods
DOMinant workshop - Molde - September 20-22 2009
Trends in industry
• Data availability, visibility and information sharing
• Increased cooperation along the supply chain• Increased globalization• Acquisitions and mergers• Extended transport contracts, 3PL, 4PL• A number of DSSs under development• Environmental focus
DOMinant workshop - Molde - September 20-22 2009
Trends in the research literature
• Richer models• Uncertainty and risk• Industrial cases• Advanced heuristics
DOMinant workshop - Molde - September 20-22 2009
Future research directions
• Richer models and integrated systems• Focus on uncertainty, robustness and
flexibility• From cost minimization to profit maximization• Better benchmarks