Richard Lipka

Design of a Component-Based Simulation Framework for Component Testing Using SpringDM

In this paper, we present the design of a simulation framework aimed to support testing of real components in a simulation environment. This enables thorough tests of software components without the need to create their models.

SimCo – Hybrid Simulator for Testing of Component Based Applications

Testing of component-based applications is important in order to ensure that third-party components do not compromise the functionality or properties of the whole system. However, thorough testing of functionality, behaviour and extra-functional properties is a tedious and time consuming process. In this paper we present an approach to discrete event simulation testing of components and component sets. Its unique feature is the ability to execute a mixture of real, implemented components and simulated mock-ups of the remaining parts of the application. Together, this approach allows faster testing on a wide scale of different inputs for tested components. At the same time, the use of actual components increases the confidence in the simulation test results. The approach has been implemented using the OSGi platform in the form of the SimCo framework and toolset, for which the key architectural considerations are discussed together with a short case study illustrating its usage.

A Method for Semi-automated Generation of Test Scenarios Based on Use Cases

Use cases are a widely accepted way to define application functionality. They can therefore form a solid basis for testing the correct functionality and quality of service of a developed application. In this paper, we describe a method for semi-automated generation of test scenarios for simulation testing of software components and component-based applications. It derives the scenarios from the use cases written in natural language, enriched with annotations that allow us to connect the specification with the source code of the application. This helps us to generate a sequence of method invocations within the tested application that forms a testing scenario. The achieved benefit is that the utilization of use cases not only helps to keep tests directly related to the original requirements of the application, it also makes it possible to easily generate new test cases when the requirements on the application change.

Implementation of Traffic Lights in JUTS

This paper describes behaviour and implementation of light signalization in urban traffic simulator JUTS. Traffic lights are necessary part of urban traffic network and they allow using JUTS in experiments dealing with impact of time plans to traffic situation.

Analysis of fitness function of genetic algorithm for road traffic network division

In this paper, the analysis of the fitness function of a genetic algorithm is discussed. This genetic algorithm is used by a method for the road traffic network division. The division of the road traffic network into a number of sub-networks is a part of necessary preparations for a distributed road traffic simulation. The fitness function consists of two parts reflecting two important issues of the road traffic network division - the load-balancing of the resulting sub-networks and the minimization of the number of divided traffic lanes. During the optimization and refactoring of the division method, it was discovered that the fitness function of the genetic algorithm is flawed, but gives better results than a new (repaired) fitness function. Hence, the working of the original fitness function was analyzed and the new fitness function was adjusted to give similar or better results than the original fitness function.

Verification of SimCo — Simulation tool for testing of component-based application

In software testing, most attention is given to verify the functionality of produced software. However, the quality of services (such as time to response or amount of produced data) is also important, especially in world of mobile or embedded devices. These properties are often estimated by using models of components, instead of testing the components from real applications. We have developed a simulation-based, non-invasive tool for testing and verifying characteristics of components or parts of component applications. In this paper, we present our approach to testing these properties and verification of functionality of our tool. Our approach allows testing of components without changing their implementation, as a black box objects. The results of our simulation tool are compared with results obtained from self-measuring application in order to validate the functionality of our simulation.

Analysis and Optimization of Fitness Function of Genetic Algorithm for Road Traffic Network Division

In this paper, the analysis and the optimization of a fitness function of a genetic algorithm for the road traffic network division are discussed. We explain why an original flawed fitness function gave better results than a new fitness function with the flaws removed. We also describe the new penalizing fitness function, which gives better results than the former two, and its optimization, which leads to a substantial reduction of the computation time. The comparison of the results of the particular fitness functions and their performance is also part of this paper.

Automated Generator for Complex and Realistic Test Data

Some type of tests especially functional tests and stress test requires a large amount of realistic test data. We are proposing a tool that uses random number generators in order to create such data. The tool is designed to be able to set simple attributes of any Java object and in many cases also to create complex structures when objects are connected via references. Random values are created according to the rules that are added to the class definition in form of annotation to each attribute. Using this tool simplifies creating of tests, as the tester do not need a detailed knowledge of data structure. The specification of expected values is delegated to the designer of data model and becomes the part of the model. Furthermore, as the data objects are created at runtime, using reflection, the tests do not have to be changed when data carrying objects are modified.

Antipatterns causing memory bloat: A case study

Java is one of the languages that are popular for high abstraction and automatic memory management. As in other object-oriented languages, Javas objects can easily represent a domain model of an application. While it has a positive impact on the design, implementation and maintenance of applications, there are drawbacks as well. One of them is a relatively high memory overhead to manage objects. In this work, we show our experience with searching for this problem in an application that we refactored to use less memory. Although the application was relatively well designed with no memory leaks, it required such a big amount of memory that for large data the application was not usable in reality. We did three relatively simple improvements: we reduced the usage of Java Collections, removed unnecessary object instances, and simplified the domain model, which reduced memory needs up to 88% and made the application better usable and even faster. This work is a case-study reporting results. Moreover, the employed ideas are formulated as a set of antipatterns, which may be used for other applications.

Interface-based semi-automated testing of software components

The component-based software development enables to construct applications from reusable components providing particular functionalities and simplifies application evolution. To ensure the correct functioning of a given component-based application and its preservation across evolution steps, it is necessary to test not only the functional properties of the individual components but also the correctness of their mutual interactions and cooperation. This is complicated by the fact that third-party components often come without source code and/or documentation of functional and interaction properties. In this paper, we describe an approach for performing rigorous semi-automated testing of software components with unavailable source code. Utilizing an automated analysis of the component interfaces, scenarios invoking methods with generated parameter values are created. When they are performed on a stable application version and their runtime effects (component interactions) are recorded, the resulting scenarios with recorded effects can be used for accurate regression testing of newly installed versions of selected components. Our experiences with a prototype implementation show that the approach has acceptable demands on manual work and computational resources.