Wonhong Nam

Symbolic compositional verification by learning assumptions

The verification problem for a system consisting of components can be decomposed into simpler subproblems for the components using assume-guarantee reasoning. However, such compositional reasoning requires user guidance to identify appropriate assumptions for components. In this paper, we propose an automated solution for discovering assumptions based on the L* algorithm for active learning of regular languages. We present a symbolic implementation of the learning algorithm, and incorporate it in the model checker NuSMV. Our experiments demonstrate significant savings in the computational requirements of symbolic model checking.

Graph Theoretic Topological Analysis of Web Service Networks

Using graph theory, we analyze the topological landscape of web service networks formed by real-world data set, either downloaded from web service repositories or crawled by a search engine. We first propose a flexible framework to study syntactic web service matchmaking in a unified manner. Under the framework, then, the data set is analyzed from diverse perspectives and granularity. By and large, the data set is shown to exhibit small world network well and power-law-like distribution to some extent. Finally, using random graph theory, we demonstrate how to accurately estimate the size of the giant component of such web service networks.

Formal specifications and analysis of the computer-assisted resuscitation algorithm (CARA) Infusion Pump Control System

Reliability of medical devices such as the CARA Infusion Pump Control System is of extreme importance given that these devices are being used on patients in critical condition. The Infusion Pump Control System includes embedded processors and accompanying embedded software for monitoring as well as controlling sensors and actuators that allow the embedded systems to interact with their environments. This nature of the Infusion Pump Control System adds to the complexity of assuring the reliability of the total system. The traditional methods of developing embedded systems are inadequate for such safety-critical devices. In this paper, we study the application of formal methods to the requirements capture and analysis of the Infusion Pump Control System. Our approach consists of two phases. The first phase is to convert the informal design requirements into a set of reference specifications using a formal system, in this case EFSMs (Extended Finite State Machines). The second phase is to translate the reference specifications to the tools supporting formal analysis, such as SCR and Hermes. This allows us to conclude properties of the reference specifications. Our research goal is to develop a framework and methodology for the integrated use of formal methods in the development of embedded medical systems that require high assurance and confidence .

Automatic symbolic compositional verification by learning assumptions

Compositional reasoning aims to improve scalability of verification tools by reducing the original verification task into subproblems. The simplification is typically based on assume-guarantee reasoning principles, and requires user guidance to identify appropriate assumptions for components. In this paper, we propose a fully automated approach to compositional reasoning that consists of automated decomposition using a hypergraph partitioning algorithm for balanced clustering of variables, and discovering assumptions using the L* algorithm for active learning of regular languages. We present a symbolic implementation of the learning algorithm, and incorporate it in the model checker NuSmv. In some cases, our experiments demonstrate significant savings in the computational requirements of symbolic model checking.

Learning-based symbolic assume-guarantee reasoning with automatic decomposition

Compositional reasoning aims to improve scalability of verification tools by reducing the original verification task into subproblems. The simplification is typically based on the assume-guarantee reasoning principles, and requires decomposing the system into components as well as identifying adequate environment assumptions for components. One recent approach to automatic derivation of adequate assumptions is based on the L* algorithm for active learning of regular languages. In this paper, we present a fully automatic approach to compositional reasoning by automating the decomposition step using an algorithm for hypergraph partitioning for balanced clustering of variables. We also propose heuristic improvements to the assumption identification phase. We report on an implementation based on NuSMV, and experiments that study the effectiveness of automatic decomposition and the overall savings in the computational requirements of symbolic model checking.

On the computational complexity of behavioral description-based web service composition

The behavioral description-based Web Service Composition (WSC) problem deals with the automatic construction of a coordinator web service that controls a set of web services to reach the goal states. Despite its importance and implications, very few studies exist on the computational complexities of the WSC problem. In this paper, to address this problem, we present four novel theoretical findings on the WSC problem: (1) solving the composition problem of deterministic web services for a restricted case (when the coordinator web service has complete information about the states of all web services) is PSPACE-complete; (2) solving the composition problem of deterministic web services for a general case (when the coordinator web service has incomplete information about the states of web services) is EXPSPACE-complete; (3) solving the composition problem of non-deterministic web services on complete information is EXP-complete and (4) solving the composition problem of non-deterministic web services on incomplete information (which is the most general case) is 2-EXP-complete. These findings suggest that more efforts to devise efficient approximation solutions to the WSC problem is needed.

QoS-Driven Web Service Composition Using Learning-Based Depth First Search

The goal of the Web Service Composition (WSC) problem is to find an optimal composition of web services to satisfy a given request using their syntactic and/or semantic features. In this paper, in particular, we study the Quality of Services (QoS)-driven WSC problem to optimize service quality criteria, e.g., response time and/or throughput. We propose a novel solution based on Learning-based Depth First Search (LDFS). Given a set of web service descriptions including QoS information and a requirement web service, we reduce the QoS-driven WSC problem into a planning problem on a state-transition system. We then find the optimal solution for the problem using a dynamic programming based on LDFS which recently has shown a promising result.

Semantic web service composition via model checking techniques

The web service composition problem aims to find an optimal composition of web services to satisfy a given request by using their syntactic and/or semantic features when no single service satisfies it. In particular, the semantics of services helps a composition engine identify more correct, complete and optimal candidates as a solution. In this paper, we study the web service composition problem considering semantic aspects, i.e., exploiting the semantic relationship between parameters of web services. Given a set of web service descriptions, their semantic information and a requirement web service, we find the optimal composition that contains the shortest path of semantically well connected web services which satisfies the requirement. Our techniques are based on semantic matchmaking and three model checking techniques such as Boolean satisfiability solving, symbolic model checking, and abstraction and refinement technique. In the experiment, our proposal efficiently identifies optimal compositions of web services.

Efficient anytime algorithm for large-scale QoS-aware web service composition

The QoS-aware web service composition WSC problem aims at the fully automatic construction of a composite web service with the optimal accumulated QoS value. It is, however, intractable to solve the QoS-aware WSC problem for large scale instances since the problem corresponds to a global optimisation problem. That is, in the real world, traditional algorithms can require significant amount of time to finally find the optimal solution, and such an unexpected long delay is unfavourable to users. In this paper, we propose a novel anytime algorithm using dynamic beam widths for the QoS-aware WSC problem. Our algorithm generates early sub-optimal solutions and keeps improving the quality of the solution along with the execution time, up to the optimal solution if a client allows enough time. We empirically validate that our algorithm can identify composite web services with high quality much earlier than an optimal algorithm and the beam stack search.

Anytime algorithm for QoS web service composition

The QoS-aware web service composition (WSC) problem aims at the automatic construction of a composite web service with the optimal accumulated QoS value. It is, however, intractable to solve the QoS-aware WSC problem for large scale instances, since the problem corresponds to a global optimization problem. In this paper, we propose a novel anytime algorithm for the QoS-aware WSC problem to identify composite web services with high quality much earlier than an optimal algorithm and the beam stack search [3].