The general research goals will be to develop scheduling algorithms, to provide provable performance guarantees, and to explore lower bounds defining the limits of achievable performance. The performance of an algorithm on a given instance will be evaluated by comparing the algorithm's achieved benefit to the maximum possible benfit which could have been achieved on the same instance. The ratio between teh achievable benefit and the achieved benefit will be considered over an entire class of instances; the worst-case value of this ration is commonly called the competeitive ration of an algorithm.
The specific research goals will be two-fold. The first direction of research will be to develop additional parameterization which can be used to describe the competitiveness of schedulers with finer granularity. The second goal is to better understand how subtle changes in the underlying modelaffect the achievable performance.
