• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.14 no.2
• /
• pp.73-83
• /
• 2014

Many recent theoretical developments in the field of machine learning and control have rapidly expanded its relevance to a wide variety of applications. In particular, a variety of portfolio optimization problems have recently been considered as a promising application domain for machine learning and control methods. In highly uncertain and stochastic environments, portfolio optimization can be formulated as optimal decision-making problems, and for these types of problems, approaches based on probabilistic machine learning and control methods are particularly pertinent. In this paper, we consider probabilistic machine learning and control based solutions to a couple of portfolio optimization problems. Simulation results show that these solutions work well when applied to real financial market data.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.62 no.11
• /
• pp.1578-1583
• /
• 2013

In this paper, stochastic optimization algorithms of PSO (Particle Swarm Optimization) and APSO (Adaptive Particle Swam Optimization) are studied and compared. It is revealed that the APSO provides faster convergence and better search efficiency than the conventional PSO when they are adopted to find the global minimum of a two-dimensional function. The advantages of the APSO comes from the ability to control the inertia weight, and acceleration coefficients. To verify that the APSO is working better than the standard PSO, the design of a 10GHz microstrip patch as one of the elements of a high frequency array antenna is taken as a test-case and shows the optimized result with 5 iterations in the APSO and 28 iterations in th PSO.

• Journal of Electrical Engineering and Technology
• /
• v.12 no.2
• /
• pp.643-653
• /
• 2017

The paper presents controller optimization algorithm for a 12-pulse voltage source converter (VSC) based high voltage direct current (HVDC) system. To get an optimum algorithm, three methods namely conventional-Zeigler-Nichols, linear-golden section search (GSS) and stochastic-particle swarm optimization (PSO) are applied to control of 12 pulse VSC based HVDC system and simulation results are presented to show the best among the three. The performance results are obtained under various dynamic conditions such as load perturbation, non-linear load condition, and voltage sag, tapped load fault at points-of-common coupling (PCC) and single-line-to ground (SLG) fault at input AC mains. The conventional GSS and PSO algorithm are modified to enhance their performances under dynamic conditions. The results of this study show that modified particle swarm optimization provides the best results in terms of quick response to the dynamic conditions as compared to other optimization methods.

• Transactions of the Korean Society of Machine Tool Engineers
• /
• v.12 no.3
• /
• pp.82-88
• /
• 2003

SGA(Single Genetic Algorithm) is a heuristic global optimization method based on the natural characteristics and uses many populations and stochastic rules. Therefore SGA needs many function evaluations and takes much time for convergence. In order to solve the demerits of SGA, ${\mu}GA$(Micro-Genetic Algorithm) has recently been developed. In this study, ${\mu}GA$ which have small populations and fast convergence rate, was applied to structural optimization with discrete or integer variables such as 3, 10 and 25 bar trusses. The optimized results of ${\mu}GA$ were compared with those of SGA. Solutions of ${\mu}GA$ for structural optimization were very similar or superior to those of SGA, and faster convergence rate was obtained. From the results of examples, it is found that ${\mu}GA$ is a suitable and very efficient optimization algorithm for structural design.

• Journal of the Korean Mathematical Society
• /
• v.47 no.1
• /
• pp.17-28
• /
• 2010

In this paper we deal with linear chance-constrained optimization problems, a class of problems which naturally arise in practical applications in finance, engineering, transportation and scheduling, where decisions are made in presence of uncertainty. After giving the deterministic equivalent formulation of a linear chance-constrained optimization problem we construct a conjugate dual problem to it. Then we provide for this primal-dual pair weak sufficient conditions which ensure strong duality. In this way we generalize some results recently given in the literature. We also apply the general duality scheme to a portfolio optimization problem, a fact that allows us to derive necessary and sufficient optimality conditions for it.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1995.04a
• /
• pp.697-708
• /
• 1995

For many practical optimization problems where the system components are stochastic, the objective functions can not be represented analytically. Furthermore, many of these problems are characterized by the presence of multiple and conflicting objectives. In this research, we introduce a new algorithm through an interactive cutting plane method for solving this multi-criteria simulation optimization problem. Then a turning process is evaluated through the proposed algorithm.

• Advances in aircraft and spacecraft science
• /
• v.3 no.2
• /
• pp.211-224
• /
• 2016

The present study aims to investigate the shimmy stability behavior of a single wheeled nose landing gear system. The system is supposed to be equipped with an electromechanical actuator capable to control the shimmy vibrations. A Proportional-Integrative-Derivative (PID) controller, tuned by using the Particle Swarm Optimization (PSO) procedure, is here proposed to actively damp the shimmy vibration. Time-history results for some test cases are reported and commented. Stochastic analysis is last presented to assess the robustness of the control system.

• Journal of applied mathematics & informatics
• /
• v.30 no.5_6
• /
• pp.699-717
• /
• 2012

In this paper, we deal with the robust mixed $H_2/H_{\infty}$ guaranteed-cost control problem involving uncertain neutral stochastic distributed delay systems. More precisely, the aim of this problem is to design a robust mixed $H_2/H_{\infty}$ guaranteed-cost controller such that the close-loop system is stochastic mean-square exponentially stable, and an $H_2$ performance measure upper bound is guaranteed, for a prescribed $H_{\infty}$ attenuation level ${\gamma}$. Therefore, the fast convergence can be fulfilled and the proposed controller is more appealing in engineering practice. Based on the Lyapunov-Krasovskii functional theory, new delay-dependent sufficient criteria are proposed to guarantee the existence of a desired robust mixed $H_2/H_{\infty}$ guaranteed cost controller, which are derived in terms of linear matrix inequalities(LMIs). Furthermore, the design problem of the optimal robust mixed $H_2/H_{\infty}$ guaranteed cost controller, which minimized an $H_2$ performance measure upper bound, is transformed into a convex optimization problem with LMIs constraints. Finally, two simulation examples illustrate the design procedure and verify the expected control performance.

• Journal of the Korean Data and Information Science Society
• /
• v.28 no.6
• /
• pp.1245-1255
• /
• 2017

The online learning is a process of obtaining the solution for a given objective function where the data is accumulated in real time or in batch units. The stochastic gradient descent method is one of the most widely used for the online learning. This method is not only easy to implement, but also has good properties of the solution under the assumption that the generating model of data is homogeneous. However, the stochastic gradient method could severely mislead the online-learning when the homogeneity is actually violated. We assume that there are two heterogeneous generating models in the observation, and propose the a new stochastic gradient method that mitigate the problem of the heterogeneous models. We introduce a robust mini-batch optimization method using statistical tests and investigate the convergence radius of the solution in the proposed method. Moreover, the theoretical results are confirmed by the numerical simulations.

• Earthquakes and Structures
• /
• v.14 no.3
• /
• pp.215-227
• /
• 2018

Supplemental passive control devices are widely considered as an important tool to mitigate the dynamic response of a building under seismic excitation. Nevertheless, a systematic method for strategically placing dampers in the buildings is not prescribed in building codes and guidelines. Many deterministic and stochastic methods have been proposed by previous researchers to investigate the optimum distribution of the viscous dampers in the steel frames. However, the seismic performances of the retrofitted buildings that are under large earthquake intensity levels or near collapse state have not been evaluated by any seismic research. Recent years, an increasing number of studies utilize genetic algorithms (GA) to explore the complex engineering optimization problems. GA interfaced with nonlinear response history (NRH) analysis is considered as one of the most powerful and popular stochastic methods to deal with the nonlinear optimization problem of damper distribution. In this paper, the effectiveness and the efficiency of GA on optimizing damper distribution are first evaluated by strong ground motions associated with the collapse failure. A practical optimization framework using GA and NRH analysis is proposed for optimizing the distribution of the fluid viscous dampers within the moment resisting frames (MRF) regarding the improvements of large drifts under intensive seismic context. Both a 10-storey and a 20-storey building are involved to explore higher mode effect. A far-fault and a near-fault earthquake environment are also considered for the frames under different seismic intensity levels. To evaluate the improvements obtained from the GA optimization regarding the collapse performance of the buildings, Incremental Dynamic Analysis (IDA) is conducted and comparisons are made between the GA damper distribution and stiffness proportional damping distribution on the collapse probability of the retrofitted frames.