Tutorial: Writing An Effective Research Paper
A generic structure of a research paper can be as follows:
Abstract
1. Introduction
2. Background
2.1 Theory
2.2 Related Work
3. Research Achievements
3.1 Methodology
3.2 Proposal
3.3 Results
4. Discussion and Conclusions
References
This is a base-line that you can either apply "as-is" or customize to your requirements. In the latter case, note that the importance is in the roles fullfiled by each section. These are essential elements of a good research paper and thus should be accomodated.
The Abstract
The guidelines:
- Start with a few sentences describing the problem space. What is the problem, issue, question motivating your research work?
- This is followed by a few sentences on your proposals, problem solving ideas, approach of dealing with the issue.
- Finally, state the results of this work, your contribution and innovation.
The figure:
An example:
Text below taken from: Iterative multi-tier management information modeling, Jean-Philippe Martin-Flatin, Divesh Srivastava and Andrea Westerinen, IEEE Communications Magazine, pp. 92-99, December 2003.
[1. Problem Space: The management information models currently used in the Internet have several limitations. Some of them contain errors, are missing important features, or are difficult to understand. Second, standards bodies keep reinventing the wheel, which confuses the terminology (hence customers) and wastes precious time. Third, finding a good balance between too abstract, and overly detailed models is a tough challenge, rarely achieved in practice. Last, the learning curve of existing data models is too steep.] [2. Research Proposal: We propose to alleviate these problems by adopting a new process for designing and standardizing management information models. It is inspired by two techniques form software engineering: the iterative and incremental software development process, which addresses the shortcomings of the waterfall process usually adhered to by the IETF and DMTF; and multi-tier models, which capture different perspectives (e.g., analysis, design, and implementation) of the information model.] [3. Innovative Result: Our main innovations are management-architecture-neutral universal information models (UIMs), sharing of conceptual models by different standard bodies, and specialization of the people involved in designing the different layers of the models. Our new process takes into account a number of constraints identified in real-life environments.]
The Introduction
The guidelines:
- Start with a few general statements describing the problem, issue or research question.
- This is followed by a detailed analysis of the problem, issue or research question. It is important to motivate the reader to read on by showing that you are dealing with an interesting issue. Also, present your own motivation for doing research in this issue, show that the issue is important.
- Continue with a few statements on the suitability and appropriateness of your research direction based on the nature of the issue examined.
- Finally, describe the organization of the rest of the paper.
The figure:
An example:
Text below taken from: Iterative multi-tier management information modeling, Jean-Philippe Martin-Flatin, Divesh Srivastava and Andrea Westerinen, IEEE Communications Magazine, pp. 92-99, December 2003.
[1. Problem Overview After more than a decade of large-scale delpoyment, element management in the internet is generally considered a success. However, the recent challenge of integrating service management with network and systems management has shown that management information modeling, one of its building blocks is still faced with a number of problems.]
[2. Detailed Problem Description First, some management information models (MIMs) are simply not good enough. ... (goes on with details)
Second, working groups in standards organizations keep reinventing the wheel by redefining things their way. ... (goes on with details)
Third, finding a good balance between an overly abstract model and a model cluttered with low-level engineering details is renowned as a tough challenge ... (goes on with details)
Fourth, the learning curve of existing data models is too steep. ... (goes on with details) ]
[3. Suitability of Research Direction We propose to alleviate these problems by changing significantly the way MIMs are devised and standardized. Our new modeling process is very general and may serve as the basis for a wide range of management information modeling activities. ... (goes on with details) ]
[4. Paper Organization The remainder of this article is organized as follows. First, we summarize how management information modeling is currently performed in the IP world. ... (goes on with the rest)]
The Background
The guidelines:
- Structure this section to reflect a spherical coverage of related background information.
- Present and explain all the related theory and key terms. You need to provide all the required background information that will allow a non-expert reader to understand the next section describing your work.
- Finally compare and contrast your work with similar research works in order to clearly show the added value of your research.
The figure:
Research Achievements
The guidelines:
- Researchers give various titles to this section and you should decide the title that is most appropriate for you. A title can be in another generic form, e.g. "Proposed Approach" or it can simply be the name of your proposal or system.
- Start by clearly describing your research methodology. Include aspects such as assumptions made, tools used, accuracy of results, etc.
- This is followed by the description of your proposal and then any accompanying research results. Explain what your results mean.
The figure:
An example:
Text below taken from: Mobile Agent-based Performance Management for the Virtual Home Environment Christos Bohoris, George Pavlou and Antonio Liotta, Journal of Network and Systems Management (JNSM), vol. 11, no. 2, pp. 133-149, June 2003.
[1. Methodology In our evaluation
we are interested to highlight the performance overheads in the
managed network associated with the VHE performance monitoring task
based on several technologies available to date. In addition, we
consider a number of software metrics assessing each of the
approaches in terms of quality and efficiency. Within this scope we
have developed four performance monitoring systems with
similar
functionality based on Mobile Agents, CORBA, Java-RMI and the
Jasmin Script-MIB. ... (goes on with
details)
For our VHE performance monitoring systems the important areas of performance overheads involve the tasks of software migration as well as remote communication between the various entities. Measurements were taken for the following typical system operations: 1. The remote transmissions of a scheduled performance report and a real-time notification ... (goes on with details) 2. The software migration occurring in the mobile agent and Script-MIB-based systems. ... (goes on with details)
For these operations we have taken the following measurements: 1. Traffic measurements: ... (goes on with details) 2. Response times measurements: ... (goes on with details)
All measurements were taken using a testbed of Linux workstations with homogeneous features (Redhat 7.1, Pentium Celeron 466MHz and 64MB of RAM), connected to a 100Mbps Ethernet network. Each node is configured as a software router with Differentiated Services (DiffServ) support, as provided by the Linux kernel version 2.4.2.]
Discussion and Conclusions
The guidelines:
- Discuss all results and summarize the conclusions of your work. This is done in a high-level of abstraction, without the methodological or technnical details that accompanied your comments in the previous section.
- Close this section with statements on future research directions and work trigerred or inspired by your current research.
The figure:
What to do:
- Make sure you are familiar with all similar research and be very clear on your added research value.
- Ensure your methodology is scientifically correct and leads you to important results.
- Justify that your assumptions are acceptable and do not hinder your results.
- Create attractive and graphically informative figures whenever possible.
- When you start a paragraph, it is important that your intention is passed to the reader in the first sentence of the paragraph. In this way the reader knows immediately what the paragraph is about.
What not to do:
- Don't fall in the trap of working too much with an attractive idea, then try to discover a problem to fit it in. Work from the start upon important real-world problems, issues or research questions.
- Don't dive into low-level or technical details for too long, focus on the essence instead.
- Don't be plain with your results, especially when you have gathered lots of numbers. Explain the essence, why are these results important, what is the crucial information they reveal.