MULTI-ARMED BANDIT ALGORITHM FOR BEST SHARPE RATIO DETECTION

Sabrina Khurshid, Gourab Ghatak and Mohammed Shahid Abdulla ()
Additional contact information
Sabrina Khurshid: Indian Institute of Technology Delhi
Gourab Ghatak: Indian Institute of Technology Delhi
Mohammed Shahid Abdulla: Indian Institute of Management Kozhikode

No 557, Working papers from Indian Institute of Management Kozhikode

Abstract: The computational agent or program in a bandit problem is required to perform a maximization of the long-run reward obtained from the environment. However, this maximization may be constrained in several ways depending on the problem. For example, in ultra-reliable wireless communications, apart from maximizing the data rate (reward) - the network must ensure that the signal strength is not highly intermittent (risk). Similarly, in finance, an investment strategy depends on a low level of volatility (risk) just as much as it does on the average returns (reward). The risk-adjusted performance of a portfolio is characterized using a quantity known as the Sharpe Ratio. This is a ratio where the numerator is the average return - adjusted against any risk-free means of returns. Correspondingly, the denominator is the standard deviation of these returns over the same time period. In this paper we propose a bandit algorithm that discovers, with least experimentation, a particular financial time series that has the highest Sharpe Ratio among the many time series available.

Keywords: decision-making under uncertainty; bandit problem; financial times series; Sharpe Ratio. (search for similar items in EconPapers)
Date: 2023-03
References: Add references at CitEc
Citations:

Downloads: (external link)
https://iimk.ac.in/uploads/publications/IIMKWPS557ITS202303.pdf (application/pdf)
Our link check indicates that this URL is bad, the error code is: 403 Forbidden

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:iik:wpaper:557

Access Statistics for this paper

More papers in Working papers from Indian Institute of Management Kozhikode IIMK Campus PO, Kunnamanagalam, Kozhikode, Kerala, India -673570. Contact information at EDIRC.
Bibliographic data for series maintained by Sudheesh Kumar ().