Skip navigation

Investigation into an improved modular rule-based testing framework for business rules

Investigation into an improved modular rule-based testing framework for business rules

Wetherall, Jodie ORCID: 0000-0002-4786-5824 (2010) Investigation into an improved modular rule-based testing framework for business rules. PhD thesis, University of Greenwich.

Wetherall_Investigation_into_an_improved_2010.pdf - Published Version
Available under License Creative Commons Attribution Non-commercial No Derivatives.

Download (4MB)


Rule testing in scheduling applications is a complex and potentially costly business problem. This thesis reports the outcome of research undertaken to develop a system to describe and test scheduling rules against a set of scheduling data. The overall intention of the research was to reduce commercial scheduling costs by minimizing human domain expert interaction within the scheduling process.

This thesis reports the outcome of research initiated following a consultancy project to develop a system to test driver schedules against the legal driving rules in force in the UK and the EU. One of the greatest challenges faced was interpreting the driving rules and translating them into the chosen programming language. This part of the project took considerable effort to complete the programming, testing and debugging processes. A potential problem then arises if the Department of Transport or the European Union alter or change the driving rules. Considerable software development is likely to be required to support the new rule set.

The approach considered takes into account the need for a modular software component that can be used in not just transport scheduling systems which look at legal driving rules but may also be integrated into other systems that have the need to test temporal rules. The integration of the rule testing component into existing systems is key to making the proposed solution reusable.

The research outcome proposes an alternative approach to rule definition, similar to that of RuleML, but with the addition of rule metadata to provide the ability of describing rules of a temporal nature. The rules can be serialised and deserialised between XML (eXtensible Markup Language) and objects within an object oriented environment (in this case .NET with C#), to provide a means of transmission of the rules over a communication infrastructure. The rule objects can then be compiled into an executable software library, allowing the rules to be tested more rapidly than traditional interpreted rules. Additional support functionality is also defined to provide a means of effectively integrating the rule testing engine into existing applications.

Following the construction of a rule testing engine that has been designed to meet the given requirements, a series of tests were undertaken to determine the effectiveness of the proposed approach. This lead to the implementation of improvements in the caching of constructed work plans to further improve performance. Tests were also carried out into the application of the proposed solution within alternative scheduling domains and to analyse the difference in computational performance and memory usage across system architectures, software frameworks and operating systems, with the support of Mono.

Future work that is expected to follow on from this thesis will likely reside in investigations into the development of graphical design tools for the creation of the rules, improvements in the work plan construction algorithm, parallelisation of elements of the process to take better advantage of multi-core processors and off-loading of the rule testing process onto dedicated or generic computational processors.

Item Type: Thesis (PhD)
Additional Information:
Uncontrolled Keywords: scheduling processes, object oriented environment,
Subjects: Q Science > QA Mathematics > QA76 Computer software
Pre-2014 Departments: School of Engineering
School of Engineering > Department of Computer & Communications Engineering
Related URLs:
Last Modified: 14 Oct 2016 09:17

Actions (login required)

View Item View Item


Downloads per month over past year

View more statistics