Literature Review Howto
Mihail L. Sichitiu

This page is meant to help new students on writing literature reviews.

A thorough literature review is a prerequisite to any research work. The goal of the literature review is to avoid the reinvention of the wheel. One of the worst research scenarios is finding an interesting problem (that you can get excited about), then work hard at a solution, show that it's a great solution, write a great paper and when the review returns discover that somebody else had the same solution published n years before. Basically all the work is useless.

The only way to avoid this nightmarish scenario is to do a thorough literature review. The drawback with a thorough literature review is that it takes time. Lots of time. However, this is is a wise investment: a poorly done literature review can lead at a huge waste of time (the nightmarish scenario).

Researching the Literature

• your adviser can usually provide you with some leads and some articles to get you started.
• do a search on Google Scholar.
• check out the last 2-3 years of the major conferences in the domain. For WALAN the list of the conferences can be found on the topic selection howto page. All these conferences have web sites. Check the technical program for the title of the papers and mark any papers that even remotely resembles your topic of interest. Be aware that sometimes the same problem is known under different names (e.g., localization, positioning, geolocation, locationing, etc.).
• do a search on Citeseer.
• use the set of papers you found using the previous bullets as your starting set. Do forward and backward literature searches on those papers. This is by far the best method of increasing the set of papers related to your subject.
  • Read the related work (or previous work) section of all papers in the working set of papers. You'll usually find at least a few papers related to your subject. Include them in the working set. (This is a backward search.)
  • Do a search on Google Scholar or Citeseer for each paper in your working set. See what papers cited those papers, include them in the working set. (This is the forward search.)
  • Repeat this step until the working set of papers is unchanged.

Writing It Down

Once you found the complete set of papers (and I cannot overemphasize complete), you need to read them. Not all papers deserve the same attention. In general, the more a paper is related to your problem, the more attention it deserves.

Ideally you should be able to distinguish several classes of solutions more or less related to your problem (based on the closeness of the problem addressed in the paper to your problem, based on the particular technique used to solve the problem, etc.). Write down about these classes of solution, providing citations for each class.

After you wrote about each class (normally one-two paragraphs about each class) write a separate paragraph about each paper that addresses the same problem as you do. Be very careful what you write here. Do not to trash the papers even if you think lowly about it. Assume (this is often the case) that the authors of those papers will be reviewing your paper. (This is normal: after all they are the experts in your problem). Nobody likes to read negative comments about their papers. For each paper emphasize the difference between their and your solution.

Conclude the section by re-emphasizing what is the original contribution to the field.