Intelligent data-driven model for automated motorcycle Bill of Material costing

  • Frederico Marcos Topa e Ferreira

Student thesis: Master's Thesis

Abstract

This thesis investigates the challenges faced by KTM AG during the cost prediction of a new motorcycle Bill of Materials (BOM) in the project early stages. Three primary challenges are identified: temporal constraints, ambiguities in the BOM, and the margin of error in estimating part costs. The study delves into the contemporary landscape of digital transformation, highlighting the critical role of embracing technological advancements to optimize workflow efficiencies. Central to this optimization effort is the integration of Application Programming Interfaces, which facilitate data interchange and decision-making processes within KTM’s operational framework. It conducts a thorough examination of various cost estimation methods, categorizing them into qualitative and quantitative approaches while assessing their suitability across different project stages. In the intuitive method, text similarity algorithms are evaluated for their potential in refining cost estimations, particularly in early project stages. The analogical method is discussed, covering regression analysis aspects, while parametric cost assessment models and analytical methods like bottomup estimation are also examined. The reported findings provide solid insights to improve the workflow efficiency and cost prediction accuracy on early project stages. They allow to clearly define the correct tools and methods to be applied, providing a solid foundation to develop an intelligent data-driven model for automated motorcycle BOM costing.
Date of Award2024
Original languageEnglish
SupervisorThomas Schlechter (Supervisor)

Cite this

'