CONFERINŢĂ
Miercuri, 11 martie 2015, 12:30-13:30
sala 335, str. Teodor Mihaly, nr. 58-60 (campus FSEGA)

Improving Dynamic Service Selection in Adaptive Software Systems: Optimal Stopping vs. „Trivial Choice”

Speaker

Prof. dr. Oliver Skroch
Darmstadt University of Applied Sciences Germany

Abstract

Many adaptive software systems can dynamically recompose their service invocations at run time. Prominent application examples include e.g. voice over IP communication, web mashups, or self-healing software. Such systems can improve themselves also by opportunistically searching for suitable service alternatives on the Internet, or on open platforms in general. For this purpose, the speaker recently applied an optimal stopping algorithm yielding the best possible lower probability bound for the optimal choice within a predefined constant time frame. In this talk, further simulation experiments are presented that examine, compare and measure this theoretically best algorithmic optimization approach and its behavior against a static situation without self-adaptation, and against a simple and intuitive, but non-optimal greedy optimization approach.

About the Speaker

Dr. Oliver Skroch belongs to the younger generation of German computer scientists. He works as full professor for business informatics and foundations of computer science at the Faculty of Computer Science, Darmstadt University of Applied Sciences, Germany. Prior to his academic career he gained industrial experience in functional and managerial positions in software development and consulting world-wide for 10+ years. He earned his doctoral degree at the chair of business informatics and systems engineering, Augsburg University, Germany.