“…Considering the levels of detail ( 2)-( 4) the MCC, F-Score, precision and recall of the best performing model on each level increases with a finer level of detail. Thus, the model quality of the best regression models increases with a finer level of detail but still lower than the model quality of the best classification model, which was on the coarsest level of detail (1).…”
Section: Resultsmentioning
confidence: 92%
“…0.3 to 0.4 and F-scores of approx. 50 % to 60 %, the best models on the finest three levels of detail ( 2)-( 4) do not reach the result of the best model on the coarsest level of detail (1). Considering the levels of detail ( 2)-( 4) the MCC, F-Score, precision and recall of the best performing model on each level increases with a finer level of detail.…”
Section: Resultsmentioning
confidence: 98%
“…The detailed explanation of the operation principals including the ML-algorithms used on the four levels of detail (see Table I) is first given for the coarsest level of detail (1). Afterwards the operation principal of the levels of detail ( 2)-( 4) is explained.…”
Section: Figure 1 Concept Of the Prediction Models On Four Levels Of Detailsmentioning
confidence: 99%
“…They are challenged to assert themselves in international markets and to differentiate their products from other products available on the market in in terms of functionality, quality and price. Furthermore, the logistics performance, such as high adherence to delivery dates or short delivery and lead times, is becoming a key competitive factor [1][2][3]. A typical example for this are machine and plant manufacturers, whose products often consist of a large number of customized components to enable a tailor-made solution for the respective customer [4,5].…”
“…Considering the levels of detail ( 2)-( 4) the MCC, F-Score, precision and recall of the best performing model on each level increases with a finer level of detail. Thus, the model quality of the best regression models increases with a finer level of detail but still lower than the model quality of the best classification model, which was on the coarsest level of detail (1).…”
Section: Resultsmentioning
confidence: 92%
“…0.3 to 0.4 and F-scores of approx. 50 % to 60 %, the best models on the finest three levels of detail ( 2)-( 4) do not reach the result of the best model on the coarsest level of detail (1). Considering the levels of detail ( 2)-( 4) the MCC, F-Score, precision and recall of the best performing model on each level increases with a finer level of detail.…”
Section: Resultsmentioning
confidence: 98%
“…The detailed explanation of the operation principals including the ML-algorithms used on the four levels of detail (see Table I) is first given for the coarsest level of detail (1). Afterwards the operation principal of the levels of detail ( 2)-( 4) is explained.…”
Section: Figure 1 Concept Of the Prediction Models On Four Levels Of Detailsmentioning
confidence: 99%
“…They are challenged to assert themselves in international markets and to differentiate their products from other products available on the market in in terms of functionality, quality and price. Furthermore, the logistics performance, such as high adherence to delivery dates or short delivery and lead times, is becoming a key competitive factor [1][2][3]. A typical example for this are machine and plant manufacturers, whose products often consist of a large number of customized components to enable a tailor-made solution for the respective customer [4,5].…”
“…The research results are implemented into an overall heuristic for an economic selection of pre-emptive disruption management measures (Burggraef et al, 2017). This simulation research delivers the benefit calculation for disruption countermeasures as well as for measure combinations.…”
Section: Application Of the Research Resultsmentioning
Purpose -The purpose of this paper is to investigate the benefit of pre-emptive disruption management measures for assembly systems towards the target dimension adherence to delivery times.Design/methodology/approach -The research was conducted by creating simulation models for typical assembly systems and measuring its varying throughput times due to changes in their disruption profiles. Due to the variability of assembly systems, key influence factors were investigated and used as a foundation for the simulation setup. Additionally, a disruption profile for each simulated process was developed, using the established disruption categories material, information and capacity. The categories are described by statistical distributions, defining the interval between the disruptions and the disruption duration. By a statistical experiment plan, the effect of a reduced disruption potential onto the throughput time was investigated.Findings -Pre-emptive disruption management is beneficial, but its benefit depends on the operated assembly system and its organisation form, such as line or group assembly. Measures have on average a higher beneficial impact on group assemblies than on line assemblies. Furthermore, it was proven that the benefit, in form of better adherence to delivery times, per reduced disruption potential has a declining character and approximates a distinct maximum.Originality/value -Characterising the benefit of pre-emptive disruption management measures enables managers to use this concept in their daily production to minimise overall costs. Despite the hardly predictable influence of pre-emptive disruption measures, these research results can be implemented into a heuristic for efficiently choosing these measures.
Due to Digital Transformation, also called Industry 4.0 or the Industrial Internet of Things, the barrier for implementing data collecting technology on the shop floor has decreased dramatically in the past yearsleading to an increasingly growing amount of data from a multitude of IT systems in production companies worldwide. Despite that, the production controller still relies heavily on intrinsic knowledge and intuition for the management of disruptions in production. Thanks to advances in the fields of production control and artificial intelligence, potentials for the collected data for disruption management arise. However, in order to transform data into usable information and allow drawing conclusions for disruption management in production, the relevant data-objects, disturbances and alternative actions must be known. Thus, the decision-making can be supported, reducing the decision latency and increasing benefit of alternative actions. Therefore, the goal of this paper is to discuss the prerequisites necessary to perform a data based disruption management and the methodology itself, serving as an approach to allow companies to build a data basis, classify disruptions and alternative actions in order to improve decision making in the future.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.