A carregar...
Publicação
Scheduling multi-skilled tasks and engineers: a telecommunication company’s problem
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 1.84 MB | Adobe PDF |
Orientador(es)
Resumo(s)
O corrente projeto contribui para o agendamento de tarefas e de engenheiros com múltiplas qualificações. Este foi desenvolvido no âmbito de uma empresa de telecomunicações escocesa – Locktel Ltd – e surge da necessidade sentida na empresa de, automaticamente e com precisão, agendar tarefas e alocar engenheiros.
Para resolver o problema foram desenvolvidos dois modelos de programação linear (um misto e um binário) interligados. O primeiro agenda as tarefas e atribui-lhes um engenheiro – líder. O segundo identifica as tarefas, entre as já agendadas, que necessitam de mais engenheiros – parceiros – e atribui-os. O objetivo é maximizar as qualificações, requeridas pelas tarefas, dos engenheiros afetos a tarefas. A decisão de desenvolver dois modelos ajuda a abordar o problema de maximização identificado em situações variadas. Além disso, os parceiros só podem ser atribuídos quando os líderes já são conhecidos, sendo que cada engenheiro designado no primeiro modelo se torna um líder, no segundo modelo, e os restantes potenciais parceiros.
Ambos os modelos foram implementados num ficheiro Excel, através de linguagem VBA, e resolvidos, com métodos exatos, usando o OpenSolver. Tanto as leituras de inputs como a folha de output do programa Excel foram construídas de modo a corresponderem na sua plenitude ao layout do programa já utilizado pela Locktel, o BigChange.
Por fim, após a análise dos resultados obtidos, foi possível averiguar uma potencial melhoria das qualificações dos engenheiros escolhidos com o uso do programa desenvolvido. Adicionalmente, foi interessante observar a relação inversa entre o número de tarefas a realizar num só dia e a respetiva média das qualificações dos engenheiros escolhidos, realçando o crescente número de compromissos com o aumento no número de tarefas por agendar num só dia.
The current project contributes to the schedule of multi-skilled tasks and engineers. This project was developed in a Scottish telecommunications company framework – Locktel Ltd – and emerges from the necessity to automatically and accurately schedule tasks and assign engineers to those tasks. In order to address the problem, two linked linear programming models (one mixed and one binary) were developed. The first one schedules all tasks and assigns one engineer per task – leader. The second one identifies which tasks, within the already scheduled ones, need more engineers – mates – and assigns them. The goal is to maximise the tasks’ required qualifications of the allocated engineers. The decision to develop two models helps to address the identified maximisation problem in various situations. Moreover, mates can only be found once the leaders are already known, as every engineer assigned in the first model becomes a leader, in the second model, and the remain engineers are potential mates. The models were written in an Excel file, through VBA language, and solved by exact methods using the OpenSolver. Both inputs and the output sheets were built in order to thoroughly match the layout already used by Locktel’s program, the BigChange. Lastly, after the obtained results’ analysis, it was possible to observe a potential improvement in the chosen engineers’ qualifications with the help of the developed program. Additionally, it was interesting to see the inverse relation between the number of tasks to schedule in one day and the respective average of the chosen engineers’ qualifications, highlighting the increase of compromises that need to be done with the rise of unscheduled tasks in one day.
The current project contributes to the schedule of multi-skilled tasks and engineers. This project was developed in a Scottish telecommunications company framework – Locktel Ltd – and emerges from the necessity to automatically and accurately schedule tasks and assign engineers to those tasks. In order to address the problem, two linked linear programming models (one mixed and one binary) were developed. The first one schedules all tasks and assigns one engineer per task – leader. The second one identifies which tasks, within the already scheduled ones, need more engineers – mates – and assigns them. The goal is to maximise the tasks’ required qualifications of the allocated engineers. The decision to develop two models helps to address the identified maximisation problem in various situations. Moreover, mates can only be found once the leaders are already known, as every engineer assigned in the first model becomes a leader, in the second model, and the remain engineers are potential mates. The models were written in an Excel file, through VBA language, and solved by exact methods using the OpenSolver. Both inputs and the output sheets were built in order to thoroughly match the layout already used by Locktel’s program, the BigChange. Lastly, after the obtained results’ analysis, it was possible to observe a potential improvement in the chosen engineers’ qualifications with the help of the developed program. Additionally, it was interesting to see the inverse relation between the number of tasks to schedule in one day and the respective average of the chosen engineers’ qualifications, highlighting the increase of compromises that need to be done with the rise of unscheduled tasks in one day.
Descrição
Mestrado Bolonha em Métodos Quantitativos para a Decisão Económica e Empresarial
Palavras-chave
Technicians and Tasks Scheduling Problem Programação linear Visual Basic for Applications OpenSolver Linear Programming
Contexto Educativo
Citação
Editora
Instituto Superior de Economia e Gestão