Computing performance for large systems with Petri nets (without running out of memory)
|讲座名称||Computing performance for large systems with Petri nets (without running out of memory)|
|讲座人介绍||Ricardo J. Rodríguez (M’13) received the M.S. and Ph.D. degrees in computer science from the University of Zaragoza, Zaragoza, Spain, in 2010 and 2013, respectively, where his Ph.D. dissertation was focused on performance analysis and resource optimization in critical systems, with special interest in Petri net modeling techniques.
He was a Visiting Researcher with the School of Computer Science and Informatics, Cardiff University, Cardiff, U.K., in 2011 and 2012, and the School of Innovation, Design and Engineering, Mälardalen University, Västerås, Sweden, in 2014. He was also a Visiting Professor in the Second University of Naples, Caserta, Italy, during a three-month period in 2016. He is currently an Assistant Professor at University of Zaragoza, Zaragoza, Spain. His research interests include performability and dependability analysis, program binary analysis, and contactless cards security. Dr. Rodríguez was involved in reviewing tasks for international conferences and journals.
|讲座内容||Performance is a non-functional property of high interest in many systems. Consider for instance a manufacturing process, logistics, or an on-line shopping web site. In these scenarios, better performance is directly "translated" to higher income (e.g., more items can be dispatched in the same time, few time is needed to transport products, or more clients can be attended in a concurrent form). Many of these artificial systems can be naturally modeled as discrete event systems. Unfortunately, in real world these systems are usually large and complex, making the exact computation of performance a highly complex computational task -- mainly caused by the well-known state explosion problem. As a result, a task that requires an exhaustive state-space exploration becomes infeasible.
To avoid the state explosion problem inherent to discrete models, approaches that provide performance bounds can be used. In this talk, we review the literature about performance bound computation for Petri nets, providing insights and current challenges. We will also introduce a novel technique that (partially) resolves some of these challenges.