- author = "Lucija Planinic",
- title = "Improving automated design of dispatching rules in the unrelated machines environment",
- school = "Department of Electronics, Microelectronics, Computer and Intelligent Systems, University of Zagreb",
- year = "2025",
- address = "Zagreb",
- month = "11 " # jul,
- keywords = "genetic algorithms, genetic programming, scheduling problems, unrelated machines environment, dispatching rules, ensembles, batch scheduling, dispatching rule simplification, genetic programming representations, genetsko programiranje, problemi rasporedivanja, okruzenje nesrodnih strojeva, pravila rasporedivanja, ansambli, grupno rasporedivanje, pojednostavljenje pravila, rasporedivanja reprezentacije jedinki u genetskom programiranju",
-
URL = "
https://urn.nsk.hr/urn:nbn:hr:168:284774",
-
URL = "https://repozitorij.fer.unizg.hr/object/fer:14434/FILE0",
-
URL = "https://repozitorij.fer.unizg.hr/islandora/object/fer:14434/datastream/PDF/view",
- size = "131 pages",
-
abstract = "Scheduling is a process that involves assigning a set
of activities or tasks to one of the available limited
resources over a specific time period. The goal of
scheduling is to create a schedule that optimizes
certain user-defined criteria. Scheduling challenges
arise in various real-world scenarios, such as
manufacturing processes, airports, and computer
clusters. Many scheduling problems fall into the
NP-hard category, making them difficult to solve. As a
result, heuristic methods are frequently used to solve
scheduling problems. One widely used approach is
dispatching rules, which differ from methods that
iteratively refine schedules. Dispatching rules create
schedules incrementally by deciding which job should be
assigned to which machine at each decision point,
making them especially useful in dynamic environments
where conditions change frequently. However, designing
effective dispatching rules is complex and
time-consuming. To address this issue, genetic
programming is often used to automatically generate new
dispatching rules.
The main goal of this thesis is to improve the performance of dispatching rules generated by genetic programming. This can be achieved through different methods. One research area focuses on comparing different solution representations to determine which is the most suitable for evolving dispatching rules. The second method proposes the application of simplification procedures (reduction and pruning) to dispatching rules evolved by genetic programming to increase their interpretability and reduce redundant sub-expressions. Another direction focuses on methods for combining existing dispatching rules into ensembles to enhance their performance. Lastly, automatically generating dispatching rules with genetic programming to solve batch scheduling problems in which one machine can process multiple jobs at the same time is explored.",
-
notes = "In English
doktorski_rad_lucija_planinic_uvez-pdf-a.pdf
Supervisor: Marko Durasevic",
