Optimal strategy for time-limited sequential search

An optimal strategy is developed for maximizing the expected benefit in time limited sequential search processes with very low candidate encounter rates. We formulate a model for searches where there are k types of candidates whose benefits and encounter rates are known prior to the start of the search. The optimal strategy consists of determining the specific times during the search at which the acceptable candidate pool should be expanded by including the next lower candidate type in the pool. Results indicate that, in general, candidate types with the higher benefits dominate the optimal strategy.