ForecastingStock Control Interactions III

1
Forecasting/Stock Control Interactions III

• Tarkan Tan
• Technische Universiteit Eindhoven
• October 23, 2007
• Forecasting and Inventory Management Bridging
  the Gap
• EPSRC project Meeting - London

2
Outline

• General Observations The Role of Forecasting in
  Production/Inventory Systems
• Research Interests
• Past Research
• Ongoing Research
• Future Research

3
General Observations The Role of Forecasting in
Production/Inventory Systems

• Deterministic Demand
• Point Estimate for Future Demand
• Mathematical Programming Models
• E.g., aggregate production planning
• Lot sizing / EOQ models
• Materials requirement planning
• Coordinated replensihment
• Stochastic Models for other uncertainties
• supply
• price
• capacity
• etc
• etc

4
General Observations (contd)

• Stochastic Demand
• Demand Distribution
• Stationary
• A wide variety of models and optimality results
• Non-stationary
• State-dependent policies

5
The Gap between Forecasting and Inventory Control

• Demand distributions based on forecasting, time
  series models, Bayesian models, etc. do not
  capture the dynamic nature of the forecasting
  component(s) of the problem.
• The effects of not only the forecast of the most
  imminent period, but also the forecasts for the
  following periods could be taken into account on
  the production/inventory strategy.
• When forecasting the demand of a number of items,
  there may exist correlations among the forecasts.
• Models that take these aspects into account are
  can/do improve system performance.

6
Some attempts of using data / information /
forecast directly in inventory planning models

• Martingale Model of Forecast Evolution (Heath and
  Jackson 1991, Güllü 1997)
• Advance Demand Information
• Perfect (Hariharan and Zipkin 1995, Gallego and
  Özer 2001, Karaesmen et al. 2002)
• Imperfect (Van Donselaar et al. 2001, Zhu and
  Thonemann 2004, Tan et al. 2007)
• Other methods
• Demand modelled as an autoregressive moving
  average process (Johnson and Thompson 1975,
  Erkip, Hausman, and Nahmias 1990, Gilbert 2005)
• Edemand follows an exponential smoothing
  formula (Miller 1986)
• Bayesian model for evolving estimates of the
  demand distribution (Scarf 1959, Azoury and
  Miller 1984, Azoury 1985)

7
Multi-Echelon Inventory Systems

• Sharing forecast information forecasts
  communicated between supply chain members
• Additional concerns
• Revealing forecast updates (before firm orders)
• Forecast volatility too frequent or large
  updates
• gt manufacturer ignores revisions
• Truthful reveal?
• Forecast inflation to ensure sufficient supply
• gt manufacturer penalizes the retailer for
  unreliable forecasts by providing lower service
  levels
• gt retailers penalize suppliers that have a
  history of poor service by providing them with
  overly inflated forecasts
• Lose-lose situation! (Terwiesch et al. 2005)
• Mostly analyzed by game-theoretic models
• Price or capacity as decision variable
• Contracting issues

8
Spare Parts Inventory Control Systems (Service
Logistics)

• Growing Interest (increasing revenues, much
  higher profitability)
• Differences with "regular" inventory control
• Low, sporadic, and highly non-stationary demand
  rates, strong dependencies
• Statistical forecasting is much harder
• More of "risk management" than inventory control
• Machine up-time multi-item approach
• Further complicating factors
• High service requirements
• Various service level aggreements
• Commonalities
• Transshipment issues (lateral, multiple-mode,
  etc.)
• etc

9
Research Interests - Past Research

• Advance Demand Information
• Capacity Management
• Service Logistics / Spare Parts Management

10
Advance Demand Information

• Tan, T., Güllü, A. R., and Erkip, N. (2007),
  Modelling Imperfect Advance Demand Information
  and Analysis of Optimal Inventory Policies,
  European Journal of Operational Research, 177,
  897-923. ADI-1.ppt

11
Advance Demand Information

• Tan, T., Güllü, A. R., and Erkip, N., Employing
  Imperfect Advance Demand Information in Ordering
  and Inventory Rationing Decisions, WP 2004.
  ADI-2.pdf

12
Advance Demand Information

• Tan, T., Using Imperfect Advance Demand
  Information in Forecasting, WP 2007. ADI-3.ppt

13
Capacity Management

• Tan, T. and Alp, O., An Integrated Approach to
  Inventory and Flexible Capacity Management under
  Non-stationary Stochastic Demand and Set-up
  Costs, WP 2005. CM-1.ppt

14
Capacity Management

• Alp, O. and Tan, T. (2007), Tactical Capacity
  Management under Capacity Flexibility, IIE
  Transactions (to appear). CM-2.ppt

15
Capacity Management

• Mincsovics G., Tan T., and Alp, O., Integrated
  Capacity and Inventory Management with Capacity
  Acquisition Lead Times, WP 2006. CM-3.ppt

16
Capacity Management

• Pac, M. F., Alp, O., and Tan, T., Integrated
  Workforce Capacity and Inventory Management Under
  Temporary Labor Supply Uncertainty, WP 2007.
  CM-4.ppt

17
Service Logistics / Spare Parts Management

• Van Kooten, J. P. J. and Tan, T. The Final Order
  Problem for Repairable Spare Parts under
  Condemnation, WP 2007. SL.ppt

18
Ongoing Research

• Revisions on past research
• Minimizing maximum hazard risk in HazMat
  transportation (with Osman Alp)
• Production/inventory models with stepwise
  production costs (with Osman Alp)
• Deciding on RFID tagging levels (with Evsen
  Korkmaz)
• Capacity management under supply uncertainty
  (with Refik Güllü and Simme Douwe Flapper)
• A simple heuristic for integrated capacity and
  inventory management (with Osman Alp and Ton de
  Kok)
• Multi-echelon spare parts management under batch
  ordering in the central warehouse (with Engin
  Topan and Pelin Bayindir)

19
Future Research

• Health Care Operations Management
• Service Logistics
• Forecasting and Inventory Management Bridging
  the Gap