Maintenance scheduling, an essential part of industrial operations

Good maintenance policies and practices are important for both productivity, efficiency, safety and cost-effectiveness. Gabrijela Obradovićs dissertation deals with scheduling of aircraft component maintenance and of the maintenance workshop.

– Maintenance can be a bit neglected by companies, as many of the maintenance costs are hidden. The companies therefore often operate in a reactive mode which can be quite expensive. I have worked in an industrial project with the aerospace and defence company Saab AB. The main goal of military aircraft is to maximize operational readiness, and this can be disturbed by both planned and unplanned maintenance. Additionally, in aerospace in general, a failure might be fatal, making safety and good maintenance practice non-negotiable.

Preventive maintenance scheduling for a fleet of aircraft
Preventive maintenance scheduling for a fleet of aircraft and maintenance workshop scheduling, with operational demand as input to the model and the scheduling of component replacement and repair as output.

Gabrijela has developed an integrated and circular supply chain and maintenance operations model, where the integrated parts are optimized simultaneously. The work is all based on discrete optimization, which is one of the more difficult branches of optimization. One of the solution approaches has been to use relaxation techniques like Lagrangean relaxation to obtain lower bounds of solutions that can not be obtained in real time.

Results to help decision making

The resulting mathematical solutions covered the problem that was the starting point of the work, though some details still remain, and gave a framework for the decision makers to compare what will happen in different scenarios. For some models that proved to be computationally demanding Gabrijela developed mathematical solution methods that enabled approximate solutions within reasonable time. Though the work has been done in connection with a certain industrial project, the modelling and methodology can be used in many other applications as well.

Gabrijela did her bachelor and master’s studies in Novi Sad, Serbia, and her master’s thesis in Paris, France. She knew PhD students who studied in Sweden and read about the country, and decided that she wanted to continue her career there as well. From the positions that was offered to her, she chose the one at Chalmers. But what was it that attracted her?

Work and leisure

– Sweden has beautiful nature, and you have easy access to it. I go running all around the year and enjoy hiking in the summer and skiing in the winter, and I also go climbing though mostly indoors. But Sweden is also known for caring about work-life balance, which is quite attractive, and I think that people do try to have a healthy balance.

The research work itself has been a lot of ups and downs. It took time to understand what it really is “to do research”. As a PhD student in mathematics, you often work a lot alone and Gabrijela thinks it would have been nice to work together with somebody except the supervisors. The teaching part of the position was a very nice experience, it was rewarding to see the students understand what was explained to them. Gabrijela has mostly taught optimization master’s courses, as well as some more general courses such as statistics and linear algebra. The relation between seniors and PhD students she has found very friendly and she has also liked the international environment at Chalmers where you meet people from many different cultures.

After the thesis defence, Gabrijela will look for a job in industry, and stay in Sweden for now. She also plans for a travel to the northern parts of the country which she has hitherto not visited.

Gabrijela Obradović will defend her PhD thesis  “Mixed-integer optimization modelling for the simultaneous scheduling of component replacement and repair” on September 7 at 10.00 in lecture hall Pascal, Hörsalsvägen 1. Supervisor is Ann-Brith Strömberg, industrial advisor Kristian Lundberg.

Link to the PhD thesis

Author

Setta Aspström