Scheduling and Planning algorithms for Electronic Calendar Management

Abstract

As part of my doctorate thesis I am looking into scheduling and planning problems and systems with a focus on electronic calendar applications. Cooperating with my supervisor Associate Professor Ioannis Refanidis, I have co-developed such a system, entitled SELFPLANNER (Refanidis and Alexiadis 2007) (Refanidis and Alexiadis 2008), which can be accessed at http://selfplanner.uom.gr. The system SELFPLANNER started as my master thesis. It is a scheduling system that supports simple, interruptible and periodic activities, where the user can define their minimum and maximum durations, their temporal domain, alternative locations, ordering constraints and temporal preferences. Specifically, in regards to to the temporal domain, an original method of defining an activity’s domain was developed (Alexiadis and Refanidis 2009). This method combines the definition and application of templates with manual editing of the calendar, by the user, to specify the temporal domain. An extended version of the scheduling problem definition was published (Refanidis and Yorke-Smith 2010) and the system was extended to incorporate that model. In the extended model, the system solves a scheduling problem, where it attempts to maximize a global utility function for the particular problem instance being solved, while meeting all the constraints of the problem. An activity can be omitted from the resulting schedule without a violation of the problem’s constraints. However, doing so will result in that activity’s utility contribution not being added to the resulting solution’s utility value. Moreover, any binary preferences involving said activity will be ignored and thus provide no utility to the solution. The extended system also supports activities overlapping in time, based on the activities’ utilization value. The work resulting to the system’s extension was published in the journal Computation Intelligence (Refanidis and Alexiadis 2011). Alternative algorithms were designed and developed to solve SELFPLANNER’s scheduling problem. The first of them was based on a genetic algorithm approach. Each chromosome of the population corresponds to the set of activi-

Extracted Key Phrases

Cite this paper

@inproceedings{Alexiadis2013SchedulingAP, title={Scheduling and Planning algorithms for Electronic Calendar Management}, author={Anastasios Alexiadis}, year={2013} }