Stefan Schupp

SMT-RAT: An Open Source C++ Toolbox for Strategic and Parallel SMT Solving

During the last decade, popular SMT solvers have been extended step-by-step with a wide range of decision procedures for different theories. Some SMT solvers also support the user-defined tuning and combination of such procedures, typically via command-line options. However, configuring solvers this way is a tedious task with restricted options.

Perfect difference sets for neighbor discovery: energy efficient and fair

We present an energy efficient neighbor discovery framework that enables Linux and TinyOS based systems to discover and connect to neighbors via IEEE 802.11 and IEEE 802.15.4, which are only available sporadically. Using quorum schemes, we schedule on and off times of the wireless transmitters, to guarantee mutual discovery with minimum power given a specific latency requirement. Neighbor discovery is fundamental to intermittently connected networks, such as disruption and delay tolerant networks and optimizing it, can lead to significant overall energy savings. Using perfect difference sets, our results indicate that we reduce the latency by up to 10 times at a duty cycle of 2% compared to the state of the art. We further define and characterize our neighbor discovery scheme with respect to fairness for asymmetric energy scenarios. Using these results, we allow energy-harvesting applications to adjust neighbor discovery based on their current energy requirements as a well defined trade-off.

A Benchmark Suite for Hybrid Systems Reachability Analysis

Since about two decades, formal methods for continuous and hybrid systems enjoy increasing interest in the research community. A wide range of analysis techniques were developed and implemented in powerful tools. However, the lack of appropriate benchmarks make the testing, evaluation and comparison of those tools difficult. To support these processes and to ease exchange and repeatability, we present a manifold benchmark suite for the reachability analysis of hybrid systems. Detailed model descriptions, classification schemes, and experimental evaluations help to find the right models for a given purpose.

Current Challenges in the Verification of Hybrid Systems

Latest developments brought interesting theoretical results and powerful tools for the reachability analysis of hybrid systems. However, there are still challenging problems to be solved in order to make those technologies applicable to large-scale applications in industrial context. To support this development, in this paper we give a brief overview of available algorithms and tools, and point out some of their individual characteristics regarding various properties which are crucial for the verification of hybrid systems. We present exemplary evaluations on three benchmarks to motivate the need for further development and discuss some of the main challenges for future research in this area.

HyPro: A C++ Library of State Set Representations for Hybrid Systems Reachability Analysis

In this tool paper we introduce HyPro, our free and open-source C++ programming library, which offers implementations for the most prominent state set representations used by flowpipe-construction-based reachability analysis techniques for hybrid systems.

Divide and Conquer: Variable Set Separation in Hybrid Systems Reachability Analysis

In this paper we propose an improvement for flowpipe-construction-based reachability analysis techniques for hybrid systems. Such methods apply iterative successor computations to pave the reachable region of the state space by state sets in an over-approximative manner. As the computational costs steeply increase with the dimension, in this work we analyse the possibilities for improving scalability by dividing the search space in sub-spaces and execute reachability computations in the sub-spaces instead of the global space. We formalise such an algorithm and provide experimental evaluations to compare the efficiency as well as the precision of our sub-space search to the original search in the global space.

Efficient Dynamic Error Reduction for Hybrid Systems Reachability Analysis

To decide whether a set of states is reachable in a hybrid system, over-approximative symbolic successor computations can be used, where the symbolic representation of state sets as well as the successor computations have several parameters which determine the efficiency and the precision of the computations. Naturally, faster computations come with less precision and more spurious counterexamples. To remove a spurious counterexample, the only possibility offered by current tools is to reduce the error by re-starting the complete search with different parameters. In this paper we propose a CEGAR approach that takes as input a user-defined ordered list of search configurations, which are used to dynamically refine the search tree along potentially spurious counterexamples. Dedicated datastructures allow to extract as much useful information as possible from previous computations in order to reduce the refinement overhead.

Spread the Work: Multi-threaded Safety Analysis for Hybrid Systems

We consider a method for the bounded safety analysis of hybrid systems, whose continuous behaviour is intertwined with discrete execution steps. The method computes a tree of state sets, which together over-approximate reachability by bounded-length executions. If none of the state sets intersects with a given set of unsafe states then we have proven bounded safety. Otherwise, we iteratively repeat parts of the computations with locally refined search parameters, in order to reduce the over-approximation error.

Analyzing Hybrid Petri nets with multiple stochastic firings using HyPro

Two approaches have been presented for the analysis of HPnGs, (i) a tree-based approach and (ii) a state-space representation based on computational geometry. We present a translation of the tree-based representation into a geometric representation using the C++ library HyPro, which has been developed for the analysis of hybrid automata. This allows the representation and efficient and accurate analysis of HPnGs with multiple stochastic firings.