Over the last two decades there has been an increase in the research of artificial neural networks (ANNs) to forecasting problems. Both in theoretical and empirical works, ANNs have shown evidence of good performance, in many cases outperforming established statistical benchmarks. This thesis starts by reviewing the advances in ANNs for time series forecasting, assessing their performance in the literature, analysing the current state of the art, the modelling issues that have been solved and which are still critical for forecasting with ANNs, thereby indicating future research directions. The specification of the input vector is identified as the most crucial unresolved modelling issue for ANNs’ accuracy. Notably, there is no rigorous empirical evaluation of the multiple published input variable selection methodologies. This problem is addressed from four different perspectives. A rigorous evaluation of several published methodologies, along with new proposed variations, is performed on low frequency data, exploring which input variable selection methodologies perform best. This analysis concludes that regression based methodologies outperformed other linear and nonlinear ones. The best way to code deterministic seasonality in the inputs of the ANNs is explored, a topic overlooked in the ANN literature, and a parsimonious encoding based on seasonal indices is proposed. The effect of the frequency of the time series on specifying the inputs for ANNs for forecasting is evaluated, revealing several challenges in modelling high frequency time series and providing evidence that the performance of several input variable specification methodologies is not consistent for different data frequencies. This leads to an evaluation of methodologies to select input variables for ANNs solely for high frequency data. Regression based methodologies are found to perform best, in agreement with the evaluation on low frequency dataset, while the ranking of the remaining methodologies is found to be inconsistent for different data frequencies.

No description available.

Service-Oriented Architecture (SOA) is becoming the mainstream for providing efficient and agile business solutions that can keep up with changes demanded by the business world. The task of identifying candidate services is one of the main activities in developing software service-oriented models. The current service identification approaches exhibit some limitations, where they are either too complex to be adopted, too simple to satisfy all SOA principles, or are theoretical approaches that are far from automation. On the other side, ontologies have been introduced in a number of phases in the SOA development lifecycle, mainly in the design and implementation phases. This research is concerned with introducing an ontology-based framework to semantically generate services from business process architectures (BPA) and using a simple service identification approach that is based on the BP A for an organisation. In this research, a novel architectural framework has been introduced that is generic, extensible, BPA-driven, ontology-based and domain independent, but uses a simple and automatic service identification approach that adheres to SOA principles. We named this framework BPAOntoSOA, and its activities are structured into two main layers. The first layer is concerned with generating a business process architecture ontology, namely the BPAOnt ontology, for a particular organisation given the Riva-based business process architecture of that organisation and the set of associated business process models. The second layer is concerned with identifying the software services using the generated BP AOnt ontology for that organisation. In order to realise this framework, we developed the abstract BP AOnt ontology to provide a conceptual representation of Riva BP A elements and the associated BPM elements. We have also proposed a novel service identification approach based on the Riva BPA and utilised its simplicity (as being systematic in identifying and modelling a BPA) to identify services in a simple and straightforward manner while satisfying SOA principles. The BP AOntoSOA framework, the BP AOnt ontology and the proposed service identification approach have been evaluated using the Jordan's Cancer Care and Registration (CCR) processes case study in an attempt to assess the correctness of the service identification approach as well as the framework behaviour in general. This has revealed that: (1) the BP AOnt ontology contributes remarkably not only as a major component in the BP AOntoSOA framework, but also as a source of business knowledge describing the process architecture and associated business process models of an organisation to be extracted and reused, (2) the service identification approach is simple, automatic and conforms to SOAprinciples, (3) the introduction of the concept of RP A clusters to the Riva method in order group related architectural elements forming the basis for identifying services and, (4) a further suggested modification to the Riva method in order to provide better conformance to SOA principles and hence better alignment between BP As and the service oriented model of computing. In Conclusion, this research has contributed to align business process architecture and the service oriented model of computing through the newly introduced BPAOntoSOA framework. xix

No description available.

No description available.

No description available.

No description available.

No description available.

The objectives of the research reported in this thesis are to explore how enterprise resource planning (ERP) can be implemented successfully in Indian medium sized manufacturing firms and to understand why only some firms are able to implement ERP successfully, while others experience ERP failure. A review of existing literature addressing ERP implementation in both large companies and SMEs was undertaken to identify specific gaps in the literature. The resource-based view (RBV) of the firm, absorptive capacity and the diffusion of innovations (DOI) were identified as insightful theoretical bases for the study. A qualitative research methodology was adopted and, following a pilot study, data was collected through semi-structured, multi-respondent, in-depth interviews in nine case study firms. Data was analysed at three levels - within case, cross-case and group analysis. The empirical findings indicate that firms undertaking a big bang ERP implementation through well-known international vendors are more likely to be successful than those firms adopting a step by-step-approach with local vendors. A number of critical success factors (CSFs), consistent with existing studies of ERP implementation, were identified. This study demonstrates strong interlinkages between the identified CSFs, with top management support playing a central and enabling role amongst the other CSFs. Furthermore, the study identified high absorptive capacity as another important CSF for ERP implementation. The thesis offers a number of contributions. Firstly, it introduces the concept of absorptive capacity into a consideration of the CSFs of ERP implementation. It also provides a more in-depth understanding of the CSFs necessary for ERP implementation, how these are inter-related and demonstrates the central role of active top management involvement throughout the implementation process.

No description available.