Titulo Estágio
Optimization in the scheduling of molds manufacturing processes through Process Mining
Áreas de especialidade
Sistemas Inteligentes
Engenharia de Software
Local do Estágio
Politécnico de Leiria
Enquadramento
The rise of Industry 4.0 – the integration of Internet of Things (IoT), cloud computing, data processing and analytics into the factories infrastructure – enabled the possibility to automatically collect, transform and process data from the manufacturing floors. The organization of these data in the form of event logs (logs which include, for each record, a certain product or service identifier, a timestamp, an activity and the resources needed to perform it), allow for the application of Process Mining algorithms to discover, compare and enhance manufacturing processes.The molds industry is characterized by having ad-hoc, “spaghetti”-like manufacturing processes, meaning that no two molds are produced the same way, due not only to being different from each other (even if slightly), but also to have singular production contexts (including, for instance, resources’ availability at a certain moment, time constraints and overall activity in the factory).By applying Process Mining, we can have a notion of the real processes being executed in the factory floor, their conformance to certain required procedures, and the opportunities to enhance these processes.However, and considering the ad-hoc scenarios of molds’ production, optimizing these processes should account for not only the “ideal” sequence of activities traced by Process Mining, but also for its optimized scheduling. In fact, the resources’ availability as well as the level of activity existing in a certain moment at the factory floor can determine different sequences of activities and resource usages, even if for two similar molds.In this work, we will propose an approach for the scheduling of molds manufacturing processes, considering their context and historical data retrieved from Process Mining algorithms. With this approach, production engineers will receive support to decide which next activities should be performed for the molds being produced, as well as which optimized resource (human and/or machine) allocation should be put into place.There are important studies that focus on scheduling of multiple processes, such as the one in Choueiri et al. (2021) which uses Process Mining algorithms to extract the underlying industrial process via Petri nets. However, this article is focused on business processes with less variability and more repetitiveness, as well as manufacturing products based on production lines. In our case, we will be focused on the particular characteristics of mold production, and on the decision support needed for production engineers to deal with these production ad-hoc scenarios.ReferencesChoueiri, A.C., Portela Santos, E.A. Multi-product scheduling through process mining: bridging optimization and machine process intelligence. J Intell Manuf 32, 1649–1667 (2021). https://doi.org/10.1007/s10845-021-01767-2
Objetivo
The main objectives of this work are: i. Understand the current state-of-the-art in the use of Process Mining to optimize scheduling processes and how this can be adapted to fit the mold industry; ii. Pre-process and arrange a molds production dataset that can be used for this thesis (we already have data from a molds company); iii. Apply Process Mining techniques to discover activities’ sequences and resource dependencies and contextual information that can be used in the scheduling; iv. Apply a scheduling technique for the next molds’ production activities in the factory floor.
Plano de Trabalhos - Semestre 1
Sept. 2022 – Dec. 2022: state of the art for multi-perspective Process Mining algorithms and scheduling in ad-hoc scenarios;Nov. 2022 -Feb. 2023: Pre-process and arrange the dataset;Feb.2023: Apply Process Mining and scheduling techniques in their independent and simpler forms;
Jan. 2023 - Feb. 2023: Writing of the intermediate progress report.
Plano de Trabalhos - Semestre 2
Feb. 2023 – April. 2023: Apply scheduling techniques together with Process Mining results and current contextual information;Mar.2023 – May 2o23: Incorporate results in an integrated platform;Feb. 2023 – June 2023: Write thesis report.
Condições
Orientador
Rui Rijo, Ricardo Martinho, Carlos Grilo
rui.rijo@ipleiria.pt 📩