• A normal approximation model for safety stock optimization in a two-echelon distribution system

      Desmet, Bram; Aghezzaf, El-Houssaine; Vanmaele, Hendrik (Journal of the Operational Research Society, 2010)
      This paper presents an approximation model for the retailer replenishment lead-times in a two-echelon distribution system, and discusses its implementation for safety stock optimization in a one-warehouse and N-identical retailers system. The model assumes normality of demand and nominal lead times. It takes into account not only the averages of these parameters but also their variances. This approximation model is first tested on a two-echelon, one-warehouse and N-identical retailers system using discrete event simulation. It is then applied to optimize the safety stock in a two-echelon distribution system of a European market leader in the production and distribution of air conditioning equipment. Results of this implementation are analysed and discussed in detail.
    • Determining the use of data quality metadata (DQM) for decision making purposes and its impact on decision outcomes — An exploratory study

      Moges, Helen-Tadesse; Van Vlasselaer, Véronique; Lemahieu, Wilfried; Baesens, Bart (Decision Support Systems, 2016)
      Decision making processes and their outcomes can be affected by a number of factors. Among them, the quality of the data is critical. Poor quality data cause poor decisions. Although this fact is widely known, data quality (DQ) is still a critical issue in organizations because of the huge data volumes available in their systems. Therefore, literature suggests that communicating the DQ level of a specific data set to decision makers in the form of DQ metadata (DQM) is essential. However, the presence of DQM may overload or demand cognitive resources beyond decision makers' capacities, which can adversely impact the decision outcomes. To address this issue, we have conducted an experiment to explore the impact of DQM on decision outcomes, to identify different groups of decision makers who benefit from DQM and to explore different factors which enhance or otherwise hinder the use of DQM. Findings of a statistical analysis suggest that the use of DQM can be enhanced by data quality training or education. Decision makers with a certain level of data quality awareness used DQM more to solve a decision task than those with no data quality awareness. Moreover, those with data quality awareness reached a higher decision accuracy. However, the efficiency of decision makers suffers when DQM is used. Our suggestion would be that DQM can have a positive impact on decision outcomes if it is associated with some characteristics of decision makers, such as a high data quality knowledge. However, the results do not confirm that DQM should be included in data warehouses as a general business practice, instead organizations should first investigate the use and impact of DQM in their setting before maintaining DQM in data warehouses.
    • From one-class to two-class classification by incorporating expert knowledge: Novelty detection in human behaviour

      Oosterlinck, Dries; Benoit, Dries F.; Baecke, Philippe (European Journal of Operational Research, 2020)
      One-class classification is the standard procedure for novelty detection. Novelty detection aims to identify observations that deviate from a determined normal behaviour. Only instances of one class are known, whereas so called novelties are unlabelled. Traditional novelty detection applies methods from the field of outlier detection. These standard one-class classification approaches have limited performance in many real business cases. The traditional techniques are mainly developed for industrial problems such as machine condition monitoring. When applying these to human behaviour, the performance drops significantly. This paper proposes a method that improves existing approaches by creating semi-synthetic novelties in order to have labelled data for the two classes. Expert knowledge is incorporated in the initial phase of this data generation process. The method was deployed on a real-life test case where the goal was to detect fraudulent subscriptions to a telecom family plan. This research demonstrates that the two-class expert model outperforms a one-class model on the semi-synthetic dataset. In a next step the model was validated on a real dataset. A fraud detection team of the company manually checked the top predicted novelties. The results show that incorporating expert knowledge to transform a one-class problem into a two-class problem is a valuable method.
    • Multi-mode schedule optimisation for incentivised projects

      Kerkhove, Louis-Philippe; Vanhoucke, Mario (Computers and Industrial Engineering, 2020)
      This research presents a novel quantitative methodology to optimise the scheduling of subcontracted projects from the perspective of the contractor. Specifically, the scenario where the contractor’s remuneration is performance dependent is investigated. Based on the incentive methodology introduced by Kerkhove and Vanhoucke (2016), a novel mixed integer programming formulation as well as a greedy local search heuristic to solve the contractor’s problem are presented and tested in a computational experiment. For this experiment, a database containing 3,150 contract-project combinations with diverse structures has been created. The results from this experiment demonstrate the efficiency of the MIP formulation even for larger problem instances, as well as the influence of the project and contract structure on the contractor’s earnings.
    • On the resource renting problem with overtime

      Kerkhove, Louis-Philippe; Vanhoucke, Mario; Maenhout, Broos (Computers & Industrial Engineering, 2017)
      In this paper the Resource Renting Problem with Overtime (RRP/overtime) is presented. The RRP/overtime is a new problem in which the assumptions of the basic RRP are combined with the possibility to schedule (parts of) activities during overtime. The addition of this extension increases the applicability of the RRP to real world problems. This paper also presents a solution technique for this extension of the resource renting problem. The solution procedure uses a scatter search heuristic to optimize a priority list, which is then in turn used by a schedule generation scheme (PatSGS). A variation on this schedule generation scheme is also used in dedicated local search procedures. The third contribution of this research is a new lower bound for the RRP/overtime problem, which is used to evaluate the results of the proposed heuristic solution method.
    • Safety Stock Optimization in Two-Echelon Assembly systems: Normal Approximation Models

      Desmet, Bram; Aghezzaf, El-Houssaine; Vanmaele, Hendrik (International Journal of Production Research, 2010)
      This paper tackles the problem of optimising safety stocks in a two-echelon assembly system. It presents and discusses several approximation models for the assembly lead-time under the assumption of normality of the assembly demand and normality of components’ nominal lead times. These approximation models are subsequently used to optimise safety stocks throughout a two-echelon assembly system. They are then tested on a particular two-echelon N-identical component assembly system. The obtained results are compared with the results of a discrete event simulation. Finally, it is shown that lead-times and safety stock results already obtained for a two-echelon distribution system can also be derived without difficulty from those of two-echelon assembly systems.
    • Strategic venture partner selection for collaborative innovation in production systems: A decision support system-based approach

      Hacklin, Fredrik; Marxt, Christian; Fahrni, F. (International Journal of Production Economics, 2006)
      Collaborative innovation and product development projects can be regarded as an emerging challenge in innovation management, being partly reflected by the currently observable industry demand for support from strategic planning tools serving this purpose. A software tool for providing operationalized decision support has been developed, based on previous research in the area of collaborative innovation success factors. By applying experience of coaching a public research organization in finding a suitable partner for joining forces in the research and development of a future energy technology, this paper presents a solution taking into account the given competitive restrictions. Being designed for usage within a coaching framework, the tool provides a multi-perspective and interactive overview of potential venture partners to the decision-makers. Furthermore, the adoption of an integrated partner selection process for supporting technology-intensive organizations in their preparation and implementation of successful collaborative ventures is suggested.