87. Optimizing schedules by the aspect of fairness
Invited abstract in session WE-2: Combinatorial Optimization I, stream Discrete Optimization.
Wednesday, 14:45 - 16:15Room: C 103
Authors (first author is the speaker)
| 1. | Martin Rónai-Kovács
|
| 2. | Szilvia Jáhn-Erdős
|
| Department of Automation and Applied Informatics, Budapest University of Technology and Economics |
Abstract
Nowadays, schedules appear at many points in our lives, and their fairness is increasingly important. This view is reinforced when we think of schedules not just as our daily schedules, our work assignments or our public transport timetables, but also as the decision support systems that help us, for example, in a loan decision. In all these cases, it is important to base solutions only on information that is relevant when making the schedule, and it is also important to be able to point out when an unfair decision is made.
Given the diversity of scheduling problems, it requires great care to create a computation method that works for all problems. One such is entity-based fairness, which can abstract away schedules sufficiently to objectively determine the degree of fairness. However, if we want to improve the schedule, we need to consider how to use this result.
The primary question to be examined is when a change in the schedule counts as an improvement, and to what extent it results in an improvement. In addition, I examine whether an optimal improvement can be identified in each case and what the necessary conditions for this are. I define a method to propose a change to the schedule.
Keywords
- Analysis and engineering of optimization algorithms
Status: accepted
Back to the list of papers