Thesis: Forex Market Trading Strategy Based on Deep Reinforcement Learning
Using heuristic optimization and deep learning for financial time series forecasting in cryptocurrency mining and blockchain applications.
Implementation of Proximal Policy Optimization for multi-agent systems in algorithmic trading.
Leveraging transformer architectures and Time2Vec for cryptocurrency price prediction.
A simulation study on reinforcement learning approaches for optimal control and data-driven control systems.
Forex Market Trading Strategy Based on Deep Reinforcement Learning
Developed and implemented novel deep reinforcement learning algorithms for algorithmic trading in foreign exchange markets. Led to multiple publications and demonstrated practical applications of RL in financial systems.
Developed DQN agent to optimize heliostat positioning for maximizing power plant efficiency in real-time solar tracking applications.
Trained an agent using PyTorch to play ping-pong using visual input from the playground, demonstrating computer vision integration with reinforcement learning.
Implemented LQG controller using MATLAB and Simulink for Optimal Control course with Dr. Hamid Khaloozadeh.
Developed robust H∞ control strategy for Networked Control Systems (NCSs) with uncertainties in communication and system dynamics.
Proposed homework questions and solved assigned problems in online class groups.
Graded homework and quizzes; recorded screen while coding example questions.
Designed and graded homework and quizzes.
Designed and graded homework assignments.
Recorded a full programming course and designed questions based on video concepts.
Microcontroller-based PI controller auto-tuning for next-generation water coolers developed for Sedna company. Approved as an alternative to compulsory military service by the Iranian government.
Department of Electrical Engineering
Khaje Nasir University of Technology
Department of Electrical Engineering
Khaje Nasir University of Technology