The Search for New Alternatives : How Human Agents Balance Exploration and Exploitation in Complex Tasks

Balancing exploration and exploitation in organizational learning entails decisions about whether and where to search for new alternatives. We report the results from a laboratory experiment that sheds light on how individual decision-makers balance exploration and exploitation in search tasks. We show that the decision whether to search is governed by different behavioral regularities than the decision where to search for a new alternative. Positive performance feedback tends to motivate more sustained exploration of the search space, while at the same time constraining search to the neighborhood of an existing alternative. An explanation for this apparently contradictory finding is that both decisions are influenced by different aspirations. The decision about whether to search is driven by aspirations about what is attainable. Positive feedback shifts aspirations upwards, thereby prolonging exploration. In contrast, where to search is governed by a simple aspiration of performance improvement, and positive feedback tends to promote exploitation by narrowing down refinement search to a particular region of the search space. We discuss the implications of our research for organizational learning and computational models of adaptation