Managing Multiple Research Projects

Alejandro Francetich

No 516, Working Papers from IGIER (Innocenzo Gasparini Institute for Economic Research), Bocconi University

Abstract: A decision maker can experiment on up to two alternatives simultaneously over time. One and only one of these alternatives can produce successes, according to a Poisson process with known arrival rate; but there is uncertainty as to which alternative is the profitable one. The decision maker only observes the outcomes of the alternatives chosen, and choosing each alternative entails a cost. Simultaneous experimentation involves higher costs but can produce more data. At the same time, since the alternatives are negatively correlated, the outcomes of either one are informative about the other. If the costs are high and she is sufficiently impatient, the decision maker never experiments on both alternatives at once. Otherwise, if she starts with a single alternative that produces no successes, she becomes gradually pessimistic and eventually takes on the other alternative while keeping the first one — despite the higher costs and the negative correlation. Keywords: Experimentation, two-armed bandits, multi-choice bandits, negatively correlated arms, Poisson process JEL Classification Numbers: D83, D90

Date: 2014
New Economics Papers: this item is included in nep-mic
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (1)

Downloads: (external link)
https://repec.unibocconi.it/igier/igi/wp/2014/516.pdf (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:igi:igierp:516

Ordering information: This working paper can be ordered from
https://repec.unibocconi.it/igier/igi/

Access Statistics for this paper

More papers in Working Papers from IGIER (Innocenzo Gasparini Institute for Economic Research), Bocconi University via Rontgen, 1 - 20136 Milano (Italy).
Bibliographic data for series maintained by ().