Jump to: Page Content, Section Navigation, Site Navigation, Site Search, Account Information, or Site Tools.
|
|
Research News
Solving certain problems in computer science is a gamble: There's no way to know in advance how long an algorithm will take to yield an answer. In this issue (see page 51), a group of researchers presents a technique for bettering the odds by imitating Wall Street: splitting the computer's investment among a "portfolio" of different programs in the hope that one of them will get lucky.
|
Science. ISSN 0036-8075 (print), 1095-9203 (online)