This paper proposes an optimal trading algorithm based on a novel application of conventional control engineering (CE). We consider a fundamental CE concept, namely, the feedback control, and apply it to algorithmic trading (AT). The concrete feedback control strategy is designed in a form of the celebrated proportional–integral–derivative (PID) model. The highly fluctuating nature of the modern financial markets has led to the adoption of a model-free realization of the generic PID framework. The control theoretical methodology we propose is combined with the advanced statistics for the historical market data. We obtain a specific log-normal probability distribution function (pdf) associated with the specific quantities associated with the available stock data. The empirical log-normal pdf mentioned above enables the necessary PID gains optimization. For this aim, we apply the data-driven optimization approaches and consider the corresponding Monte Carlo solution procedure. The optimized PID trading algorithm we propose is also studied in the Fourier analysis framework. This equivalent frequency domain representation involves a new concept in financial engineering, namely, the “stock market energy” concept. For the evaluation, we implement the proposed PID optimal trading algorithm and develop a Python-based prototype software. We finally apply the corresponding prototype software to a data set from the Binance BTC/USDT (Bitcoin/Tether) stock market. The experimental result illustrates the implementability of the proposed optimal PID trading scheme and also shows the effectiveness of the proposed CE methods in the modern AT.
Our paper presents a numerically robust control design algorithm for a class of Optimal Control Problem (OCPs). The computational approach we develop constitutes an extension of the celebrated Gauss pseudo-spectral method. The last one makes it possible to reduce the Pontryagin Maximum Principle related Hamiltonian boundary value problem to an auxiliary algebraic system. The implementable scheme we propose is numerically consistent (by implementation) and implies an effective robust control procedure for a relative small discretization grids. The pseudo-spectral methodology is used in combination with a classic differential continuation approach. The presented solution procedure can be extremely useful when a generic shooting-type solution method fails (the case of a high sensitivity or stiffness). We finally consider an illustrative example and discuss the obtained results.
Our paper deals with an advanced statistical tool for the volatility prediction problem in financial (crypto) markets. First, we consider the conventional GARCH-based volatility models. Next, we extend the corresponding GARCH-based forecasting and calculate a specific probability associated with the predicted volatility levels. As the probability evaluation is based on a stochastic model, we develop an advanced data-driven estimation of this probability. The novel statistical estimation we propose uses real market data. The obtained analytical results for the statistical probability of the levels are also discussed in the framework of the integrated volatility concept. The possible application of the established probability estimation approach to the volatility clustering problem is also mentioned. Our paper includes a concrete implementation of the proposed volatility prediction tool and considers a novel trading and volatility estimation module for crypto markets recently developed by the 1ex Trading Board group in collaboration with GoldenGate Venture. We also briefly discuss the possible application of a model combined with the data-driven volatility prediction methodology to financial risk management.
This paper is devoted to a novel formal analysis, optimizing the learning models for feedforward multilayer neural networks with hybrid structures. The proposed mathematical description replicates a specific switched-type optimal control problem (OCP). We have developed an equivalent, optimal control-based formulation of the given problem of training a hybrid feedforward multilayer neural network, to train the target mapping function constrained by the training samples. This novel formal approach makes it possible to apply some well-established optimal control techniques to design a versatile type of full connection neural networks. We next discuss the irrelevance of the necessity of Pontryagin-type optimality conditions for the construction of the obtained switched-type OCP. This fact motivated us to consider the so-called direct-solution approaches to the switched OCPs, which can be associated with the learning of hybrid neural networks. Concretely, we consider the generalized reduced-gradient algorithm in the framework of the auxiliary switched OCP.
This paper studies optimization of dynamic systems described by affine Functional Differential Equations (FDEs) involving a sup-operator. We deal with a class of FDEs-featured Optimal Control Problems (OCPs) in the presence of some additional control constraints. Our aim is to derive the first-order optimality conditions and propose an effective solution algorithm. Moreover, we are interested in applications of the resulting optimal design techniques to the Maximum Power Point Tracking (MPPT) control of solar energy plants. We develop a conceptual computational approach to the specific class of OCPs under consideration and also point possible applications of this new methodology in MPPT control.