Other Aspects

There are many other aspects that might be relevant when evaluating heuristics. Here are a few:

• Ease of Understanding. Heuristics are particularly nice when they can be explained to non-specialists. Remember that almost every project must be ``sold'' to someone else. Complicated heuristics may dazzle some people, but more likely annoy them by their lack of clarity.
• Ease of Implementation. Heuristics eventually have to be represented by computer code on some machine. A computer code that is hard to code represents real production costs and is more likely to contain programming errors that decrease the advantage of possibly improved solutions.
• Error Bounds. Some heuristics, in addition to providing a good solution, give an estimate of how far from optimal they are. For instance, the minimum spanning tree heuristic can report the value of the minimum spanning tree, and hence estimate its deviation from optimality. Based on this bound, decisions can be made as to whether to run more heuristics or do other work.
• Flexibility to change. Once a lot of work has been done to find a heuristic value, it is disappointing to redo the work if some of the numbers turn out to be wrong. The spacefilling curve heuristic can be quickly resolved if some places are a bit off. Other heuristics may require complete resolving.