Gittins Indices and Oil Exploration
将石油勘探建模为贝叶斯序贯决策问题，用吉廷斯指数确定多个勘探区域间的最优钻井策略，并评估启发式方法。
SUMMARY An earlier study by Benkherouf and Bather considered the question of when the results of oil exploration in a single area were sufficiently discouraging to render it optimal to stop drilling altogether. This stopping problem is now considerably elaborated by the inclusion of a choice between several areas in which to explore. Our oil exploration problem is modelled as a Bayesian sequential decision problem with a Markovian structure. Optimal drilling strategies are determined by Gittins indices. These indices also play an important part in the evaluation of heuristics.