Parallel search for the best alternative

Abstract

Search theoretic models, which characterize the behavior of agents and market outcomes in imperfect information environments, have been of considerable interest in a variety of economic contexts. In labor economics, such models have been used to portray the job search behavior of workers, wage determination, and to analyze labor market equilibrium. In empirical studies, they have been useful in the structural interpretation of labor market histories. There is also a vast literature in the context of consumer price search.1 Optimal sequential search strategies have also been studied in the context of research, development and exploration projects, innovation and choice of substitutable technologies, etc.2 In the search literature, two basic types of search models may be identified: random search and systematic search. Random search is typified by a scenario wherein an agent sequentially samples from the same offer distribution (offers can be prices, rewards, wages, or in general, opportunities, depending on the context), and the search decision is to decide, after obtaining a sample, whether to stop the search process or continue further search. In systematic search models, the agent faces a number of alternatives (sources, options, or projects), each with its own (possibly different) reward distribution, and thus, the agent in determining the search strategy has the choice of which alternative to explore at a point in time.3 A large body of search literature has been developed around the random search model. This assumption of random search, which has been routinely used with