In progress Demand Fulllment in Make-to-Stock Systems with Contract and Spot Customers
- ERIM PhD 2016 ESE RD CvO
Tactical level decisions for spare part management have been explored extensively. However,
due to the uncertainty of demand and supply, operational decisions such as expediting
and emergency lateral transshipment can still add value to companies. With the development
of control towers in large capital goods service companies, real time information can
be incorporated to enhance operational level decisions, but how to do it remains largely
By investigating the literature on operational aspects of spare part management and by
examining practice in companies, we identify the scope of operational decisions and figure
out future research directions. Specifically, we are interested in how to quote lead times
in a make-to-stock system with contract and noncontractual spot customers at operational
level. Furthermore, we, as well as the companies who want to utilize dynamic pricing and
advance demand information(ADI), are interested in what role dynamic pricing and ADI
would play in the demand fulfillment problem a make-to-stock system. We will use Markov
decision process models to solve the problems, and adjustments will be made for each specific
problem. We aim at finding optimal lead time quotations and the structure of optimal
solutions, which will provide managerial insights for companies, especially about new spare
parts with little historical information.
due date management; Markov decision process; dynamic pricing; advance demand information
Time frame2016 - 2020
As stated earlier, the focus of WP3 is on operational decision making, assuming that basic strategic and tactical issues have been dealt with. Our main objective therefore is to design methods and tools, and to embed them in a control tower concept, that support operational decision making in dynamically changing situations, as requested in particular by companies in response to the need to quickly adapt to rapidly changing market requirements. This leads to the following research questions:
- What are the key decisions functions of a service control tower, assuming a total supply chain and lifecycle perspective? Which decisions should be centralized and which not?
- What information is needed to timely trigger a need for operational actions, what are these actions, and how to avoid system nervousness due to overreaction?
- How to properly adapt to the various life stages of equipment (phase-in, mature, phase-out)? How to react to changing demand characteristics?
- How to operate a repair shop, given resource, parts and flexibility limitations? When to deviate from agreed working conditions/hours and decide to crash management action?
- What level of integration is needed between materials and resource planning (parts and components, engineers and tools)?
These questions reflect the idea that a sound service structure is also defined by the way it is able to handle non-routine events. That is the essential contribution of WP3, in which we will exploit methods from a variety of domains, including (stochastic) control theory, agent-based decision theory, statistical analysis, simulation and serious games.
Our approach is to use applied research with a clear focus on applications and practical use of results for the industry. Required capacity is two Postdocs or PhD students (within the time frame of three years) per involved university. Postdocs have priority because of their research experience, but the final choice will also depend on the availability of good candidates.
This research line has not been addressed as such in earlier projects like ProSeLo, but partly in MaSelMa and in own research. Nevertheless, ProSeLo’s WP1 (Smart Resources) and WP2 (Final Buy) relate to the subject. We note though that this line of research clearly connects with WP2 (Service Logistics) of MaSelMa, another innovation project in the maritime industry. The TUe and the UT are in the lead of that work package, so it makes sense to bring the UT in the lead of the of the current WP3 (with the TUe leading WP1).