Initial meeting, EPSRC CSP

1
Forecasting stock control interactions a
simulation intensive investigation
Aris A. Syntetos and Zied M. Babai CORAS -
University of Salford
2
Outline
EPSRC project
1
Forecasting and stock control
2
Current investigations preliminary results
3
Conclusions and further work
4
3
EPSRC project

• On the Development of Theory-Informed
  Operationalised Definitions of Demand Patterns.
• (FOCUS ON INTERMITTENT DEMANDS) OBJECTIVES
• To identify, through analysing the interaction
  between forecasting and stock control, the key
  factors that influence the performance of the
  total system
• To propose theoretically coherent demand
  categorisation rules for both forecasting and
  stock control purposes
• To test the empirical validity and utility of
  the theoretical results on large sets of real
  world data
• To provide a set of recommendations for
  industrial applications.

4
Methodology

• Positivistic methodology ? Development of
  universally applicable categorisation solutions
• However, due to the complexity of the problem,
  the research strategy cannot be purely
  hypothetico-deductive
• Established theory is applied to empirical data
  with the objective of identifying issues that are
  subsequently incorporated/reflected back to the
  theory. Knowledge is then deduced again and final
  recommendations/ conclusions will be made.
• Semi-deductive research strategy (theory-data
  loops) - a very well-framed simulation-intensive
  exploratory investigation.

5
Industrial collaborators
Brother International, UK
Computer Science Corporation
Valves Instrument Plus, Ltd
6
Forecasting and stock control
Estimate the lead-time demand
An appropriate demand forecasting
method (Parametric and Non-parametric methods)
1st step
An appropriate inventory control
policy (Continuous / Periodic review
policies) (Service level / Cost minimisation)
Compute the parameters of the stock control policy
2nd step
7
Demand forecasting methods

• Parametric Methods
• Known distribution is assumed (eg Poisson,
  Negative Binomial)
• Distribution parameters must be estimated
• Examples MA, SES, Crostons method
• Non-Parametric Methods
• No particular distribution is assumed
• It is assumed that distribution observed in the
  past persists into the future
• Examples Bootstrapping methods

8
Stock control methods

• Typically periodic review policies are used for
  intermittent demand items
• (T,S) and (T,s,S) policies.
• (T,S) policy Review inventory position every T
  periods and order enough to bring up to the
  order-up-to-level S
• (T,s,S) policy Inventory position dropping to
  the re-oder point s triggers a new order
• Comments on the methods
• (T,S) is very simple and performs well for low
  ordering costs
• (T,s,S) induces lower costs but the parameters
  are more complex to compute
• ?Some heuristics have been proposed to compute
  these parameters
• (Require only knowledge of mean
  and variance of the demand)

9
Current investigations

• Investigation on parametric forecasting methods
• Collaboration with Nezih Altay (University of
  Richmond, Virginia)
• Investigation on non-parametric methods
• Collaboration with John Boylan (Buckinghamshire
  New University)
• Investigation and comparison of stock control
  methods
• Collaboration with Richard Marett (Multipart)
  and Yves Dallerry (Ecole Centrale, Paris), IJPR
• A new approach for the stock control of
  intermittent demand items
• Collaboration with Ruud Teunter (Lancaster),
  JORS, EJOR
• Demand classification related issues
• Collaboration with Mark Keyes (Brother
  International), IMA

10
1 of 5 Investigation on parametric forecasting
methods

• Which distribution should be hypothesised to
  represent the demand?
• Which estimator to choose in order to forecast
  the demand?
• Limited empirical work has been conducted on
• Comparing different demand estimators
• Assessing the fit of demand distributions
• Current work
• Empirical investigation to test the statistical
  goodness-of-fit of many distributions on large
  intermittent demand datasets
• The impact of the distributional assumptions on
  stock control

11
Investigation on parametric methods

• Goodness-of-Fit results (experimentation on
  4,588 SKUs from US Navy)

Poisson distribution
Negative Binomial distribution
Normal distribution
Gamma distribution
12
2 of 5 Investigation on non-parametric methods

• Investigate and compare non-parametric
  (bootstrapping) methods
• Efrons bootstrapping Approach
• Porras and Dekkers bootstrapping Approach
• Willemains bootstrapping Approach
• Compare parametric and non-parametric methods on
  stock control performance
• Empirical results (experimentation on 1,308 SKUs
  from RAF,UK)
• Considerable cost reductions achieved by
  employing the parametric approach
• Better CSL achieved by employing the
  non-parametric approach

13
3 of 5 Investigation and comparison of stock
control methods

• Comparison of stock control methods for
  intermittent demand items
• (T,S) method
• Power Approximation (Ehrhardt and Mosier, 1984)
• Normal Approximation (Wagner, 1975)
• Naddor Heuristic (Naddor, 1975)
• Development of categorisation rules for
  inventory control purposes (experimentation on
  5,000 SKUs from RAF, UK)
• Empirical Results
• Naddors heuristic is overall the best
  performing heuristic when cost is considered
• (T,S) is the worst performing one when ordering
  cost is considered
• Consideration of both cost and service level
  results in similar performances being reported
  for all thee (T,s,S) heuristics.
• Implementation related considerations imply that
  the Power Approximation is the preferred one.

14
4 of 5 A new stock control approach

• Main assumption Lead time is smaller than the
  inter-demand interval, L Tm
• Estimating separately the inter-demand intervals
  and demand sizes, when demand occurs, directly
  for stock control purposes.

• Empirical investigation to compare the inventory
  performance of the new approach to the classical
  one (experimentation on 2,455 SKUs from the RAF,
  UK)

15
A new stock control approach

• Preliminary results
• Considerable cost reductions achieved by
  employing the new approach. The cost reductions
  range (across all SKUs) from 14 to 22
• Almost no penalty in service levels
• Extensions
• Further work is about to be submitted for peer
  review on the development of a generalised
  compound Bernoulli model
• Theoretical developments for both cost and
  service level constraints

16
5 of 5 Demand classification

• Demand categorisation in a European spare parts
  Logistics network
• In collaboration with Brother International, UK
• Typical ABC classifications
• An opportunity for considering pertinent
  qualitative issues and large scale applications
• Demonstration of the tremendous scope for
  improving real world systems
• Next steps to involve the application of
  theoretically sound solutions

17
Conclusions and further work

• Project funded by the EPSRC, UK
• Simulation intensive investigation that has been
  evolved around 5 areas
• Parametric forecasting methods
• Non-parametric methods
• Stock control methods
• Integrated forecasting stock control solutions
• Further insights into categorisation related
  issues
• We have already started reflecting pertinent
  issues identified through our empirical
  investigations into theoretical developments
• Exciting and very challenging second year of the
  project attempt to synthesise our findings into
  robust, theoretically sound, inventory management
  solutions.

18
Thank you very much
Questions ? http//www.mams.salford.ac
.uk/CORAS/Projects/Forecasting/