|
Abstracts of the seminar papers
Joaquim Gromicho (ORTEC, P.O. Box 490, 2800 AL, Gouda)
Short bio
A substantial part of our decisions are taken online. In fact, the different aspects of a problem reveal themselves during or after the process of deciding. Moreover, we often need to decide fast, just after acquiring information, and this is what we call real-time. A route planner (sometimes also called dispatcher, emphasizing the online aspects of the duty) often assigns resources to tasks before having enough information about the exact size of the load and the exact time windows of the transport. Moreover, it is often not possible to take an accurate decision with respect to the capacity to assign, because other possible cargo combinations may not have revealed themselves yet. In practice we often act as if the problem would be completely known at the moment we take a decision. Examples are insertion heuristics in a route plan. We just try to insert a new order into one of the existing trips. This is done because it is often hard to forecast new orders, or because it is hard to incorporate forecasts into the models or the software. Often the way to allow other aspects than those the model already encompasses into consideration is to let the human planner decide: the system aids the decision maker by presenting several opportunities and letting him of her choose. Bo Chen (Warwick Business School, University of Warwick, Coventry, CV4 7AL, United Kingdom) E-mail: b.chen@warwick.ac.uk Short bio
On-line methods are to address the issue of “decision making in the absence of complete information”. In this talk, we give an introduction to a standard methodology for the analysis of online decision-making, namely competitive analysis of online algorithms. An online algorithm, in contrast to a traditional algorithm, must generate output without knowledge of the entire input. Competitive analysis of online algorithms provides a framework within which real-time problems can be analyzed and solved. Peter van Tooren (Almende, Westerstraat 50, 3016 DJ,
Rotterdam)
Short bio
The presentation will deal with different aspects of the operation deployment of typical solutions in the field, such as snapshot versus continuous planning, evolutionary modeling, restriction handling and communication. The new alternatives, such as DEAL, (Distributed Engine for Advance Logistics) and agent modeling, agility, just-in-time planning and execution will be discussed. Klaas Jan van der Bent (ANWB HV-ICT / Ordina TA
Wassenaarseweg 220, P.O. Box 93200, 2509 BA The Hague)
Short bio
Attention will be given to the project "OBIdis" with an emphasis on planning methods that are used at "ANWB Wegenwacht", the unit responsible for on site servicing of car trouble. Ana Isabel Barros and Peter van Scheepstal (TNO Physics and
Electronics Laboratory, P.O. Box 96864, 2509 JG The Hague)
Short bio
This presentation will provide an overview of on-line decision making applications at TNO-FEL. Particular attention will be devoted to the project Innovation in e-fulfillment. Selling something on the Internet is much easier than bringing the goods efficiently to the customer. Many e-businesses underestimated this problem and the costs involved. TNO works on a physical distribution concept that lowers costs and increases customer service by using interactive communication with the customer. There are two main challenges in this project, building the planning algorithms to support the innovative order acceptance process and building the information architecture to support the concept. In the new concept the customer can give preferences for price differentiated delivery time windows. Afterwards he/she will immediately be informed in which time window the delivery will take place. Such time window appointments will lead to an increase in successful deliveries and hence to a decrease in logistic costs. On the other hand, setting time windows usually leads to extra costs for the logistic service provider due to loss of flexibility. This effect is minimized by taking into account current orders, future orders, routes and distribution capacity during the order acceptance process. Used on a bigger scale this concept can lead to a breakthrough in e-commerce by making it more appealing to order on-line. |