Optimal Technical Indicator Based Trading Strategies Using Evolutionary Multi Objective Optimization Algorithms

This paper proposes a bi-objective evolutionary approach to perform technical indicator-based stock trading. The objective is to find the optimal combinations of technical indicators in order to generate buy and sell strategies such that the objective functions, namely, Sharpe ratio and Maximum Draw...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Computational economics Ročník 66; číslo 1; s. 757 - 807
Hlavní autori: Vivek, Yelleti, Prasad, P. Shanmukh Kali, Madhav, Vadlamani, Lal, Ramanuj, Ravi, Vadlamani
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: Dordrecht Springer Nature B.V 01.07.2025
Predmet:
Age
Algorithms
Business cycles
Costs
Decomposition
Economic fluctuations
Evolutionary algorithms
Genetic algorithms
Genetics
Geometry
Indicators
Investments
Investors
Literature reviews
Macroeconomics
Objectives
Optimization
Popularity
Preferences
Sorting algorithms
Stock prices
Strategies
ISSN:0927-7099, 1572-9974
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:This paper proposes a bi-objective evolutionary approach to perform technical indicator-based stock trading. The objective is to find the optimal combinations of technical indicators in order to generate buy and sell strategies such that the objective functions, namely, Sharpe ratio and Maximum Drawdown, are maximized and minimized, respectively. In this study, Adaptive geometry-based MOEA (AGE-MOEA) and AGE-MOEA2 are proposed to accomplish the optimization owing to their popularity and power. This study incorporates a rolling-window-based approach (two years of training followed by a year for testing), and thus, the results of the approach seem to be considerably better in stable periods without major economic fluctuations. For the baseline comparison purpose, we employ Non-dominated sorting genetic algorithm-II (NSGA-II), Multi-objective evolutionary algorithm based on decomposition (MOEA/D) too for the problem. Further, we incorporate the transaction cost and domain expertise in the whole modeling approach. It is observed that AGE-MOEA turned out to be the best in 6 out of 11 time horizons by devising a better optimal strategy. However, MOEA/D selected less number of indicators in most of the buy strategy cases and stood first in terms of interpretability. The same observation is noticed with AGE-MOEA in sell strategy cases.
Bibliografia:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0927-7099
1572-9974
DOI:10.1007/s10614-024-10701-6