Effects of Decision Models on Dynamic Multi-objective Optimization Algorithms for Financial Markets

Maximizing profit in financial time series, like foreign exchange, with computational intelligence techniques is very challenging. It is even more challenging to make a decision from a multi-objective problem, like automated foreign exchange (Forex) trading. This study explores the effects of five d...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:2019 IEEE Congress on Evolutionary Computation (CEC) s. 762 - 770
Hlavní autori: Atiah, Frederick Ditliac, Helbig, Marde
Médium: Konferenčný príspevok..
Jazyk:English
Vydavateľské údaje: IEEE 01.06.2019
Predmet:
computational intelligence
Computational modeling
Currencies
DMOPSO-CD
DNSGA-II
DVEPSO
dynamic multi-objective
foreign exchange
Forex
Heuristic algorithms
Optimization
Particle swarm optimization
Sorting
stock market
Stock markets
swarm intelligence
Technical indicators
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:Maximizing profit in financial time series, like foreign exchange, with computational intelligence techniques is very challenging. It is even more challenging to make a decision from a multi-objective problem, like automated foreign exchange (Forex) trading. This study explores the effects of five decision models on three state-of-the-art dynamic multi-objective optimization algorithms namely, dynamic vector-evaluated particle swarm optimization (DVEPSO), multi-objective particle swarm optimization with crowded distance (MOPSO-CD) and dynamic non-dominated sorting genetic algorithm (DNSGA-II). A sliding window mechanism is employed over the USDZAR currency pair. The results show that each decision model generates different net profit. However, gray relational analysis (GRA) and objective sum (SUM) consistently performed better across all algorithms and technical indicators (relative strength index (RSI) and moving average convergence divergence (MACD)) than other decision models. Moreover, DNSGA-II was the most stable algorithm.
DOI:10.1109/CEC.2019.8790275