• IEMEK Journal of Embedded Systems and Applications
• /
• v.18 no.6
• /
• pp.319-325
• /
• 2023

In this study, we proposed a method for hyperparameter optimization in the building and training of a deep learning model designed to process point cloud data collected by a millimeter-wave radar system. The primary aim of this study is to facilitate the deployment of a baseline model in resource-constrained IoT devices. We evaluated a RadHAR baseline deep learning model trained on a public dataset composed of point clouds representing five distinct human activities. Additionally, we introduced a coarse-to-fine hyperparameter optimization procedure, showing substantial potential to enhance model efficiency without compromising predictive performance. Experimental results show the feasibility of significantly reducing model size without adversely impacting performance. Specifically, the optimized model demonstrated a 3.3% improvement in classification accuracy despite a 16.8% reduction in number of parameters compared th the baseline model. In conclusion, this research offers valuable insights for the development of deep learning models for resource-constrained IoT devices, underscoring the potential of hyperparameter optimization and model size reduction strategies. This work contributes to enhancing the practicality and usability of deep learning models in real-world environments, where high levels of accuracy and efficiency in data processing and classification tasks are required.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.23 no.8
• /
• pp.1883-1892
• /
• 1998

This paper considers the efficient design of routing path and wavelength assignment asignment in the sigle-hop WDM optical transport networks. The connecton demands between node-pairs are given and a connection must be made by only one lightpath. It is assumed that no wavelength conversion is allowed and the physical topology of the network is given. This paper proposes a method to find the routes of lightpaths and assign wavelengths to the routes, which minimizes the number of total wavelength to satisfy all connection demands. We establish a new optimization model that finds the minimum number of wavelengths. A heuristic algorithm with polynomial iterations is developed for the problem. The algorithm is implemented and applied to the netowrks with real problem size. The results of the application are compared with the commericial optimization solver, GAMS/OSL and Wauters & Demeester [8].

• The Transactions of the Korean Institute of Electrical Engineers A
• /
• v.53 no.10
• /
• pp.573-580
• /
• 2004

A new real-time constrained dispatch scheduling (CDS) is needed for TWBP. The CDS needs to be performed at every dispatch period to decide generation power of scheduling generators and amounts of scheduling load. Therefore, the CDS is not based on real generation costs but on bidding data of market participants with some constraints such as power balance, generation limits, ancillary service, and transmission line limits. This paper selects linear programming(LP) as an optimization tool for the CDS and presents effective formulae for the LP application. This paper also presents the way of minimizing the number of variables and constraints of the LP to improve real-time applicability.

• ETRI Journal
• /
• v.42 no.4
• /
• pp.562-574
• /
• 2020

Quadrature spatial modulation (QSM) utilizes the in-phase and quadrature spatial dimensions to transmit the real and imaginary parts of a single signal symbol, respectively. The improved QSM (IQSM) transmits two signal symbols per channel use through a combination of two antennas for each of the real and imaginary parts. The main contributions of this study can be summarized as follows. First, we derive an upper bound for the error performance of the IQSM. We then design constellation sets that minimize the error performance of the IQSM for several system configurations. Second, we propose a double QSM (DQSM) that transmits the real and imaginary parts of two signal symbols through any available transmit antennas. Finally, we propose a parallel IQSM (PIQSM) that splits the antenna set into equal subsets and performs IQSM within each subset using the same two signal symbols. Simulation results demonstrate that the proposed constellations significantly outperform conventional constellations. Additionally, DQSM and PIQSM provide a performance similar to that of IQSM while requiring a smaller number of transmit antennas and outperform IQSM with the same number of transmit antennas.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.34 no.2
• /
• pp.191-201
• /
• 2010

This paper presents a new global optimization algorithm based on the branch-and-bound principle using Bspline approximation techniques. It describes the algorithmic components and details on their implementation. The key components include the subdivision of a design space into mutually disjoint subspaces and the bound calculation of the subspaces, which are all established by a real-valued B-spline volume model. The proposed approach was demonstrated with various test problems to reveal computational performances such as the solution accuracy, number of function evaluations, running time, memory usage, and algorithm convergence. The results showed that the proposed algorithm is complete without using heuristics and has a good possibility for application in large-scale NP-hard optimization.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.24 no.8 s.179
• /
• pp.1940-1948
• /
• 2000

A design method for maximizing fatigue life of the deep groove ball bearing without enlarging mounting space is proposed by using a genetic algorithm. The use of gradient-based optimization methods for the design of the bearing is restricted because this design problem is characterized by the presence of discrete design variables such as the number of balls and standard ball diameter. Therefore, the design problem of rolling element bearings is a constrained discrete optimization problem. A genetic algorithm using real coding is used to efficiently find the optimum discrete design values. To effectively deal with the design constraints, a ranking method is suggested for constructing a fitness function in the genetic algorithm. Constrains for manufacturing are applied in optimization scheme. Results obtained for several 63 series deep groove ball bearings demonstrated the effectiveness of the proposed design methodology by showing that the average basic dynamic capacities of optimally designed bearings increased about 9-34% compared with the standard ones.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.24 no.8 s.179
• /
• pp.1949-1957
• /
• 2000

All the loads in the real world act dynamically on structures. Since dynamic loads are extremely difficult to handle in analysis and design, static loads are utilized with dynamic factors. The dyna mic factors are generally determined based on experiences. Therefore, the static loads can cause problems in precise analysis and design. An analytical method based on modal analysis has been proposed for the transformation of dynamic loads into equivalent static load sets. Equivalent static load sets are calculated to generate an identical displacement field in a structure with that from dynamic loads at a certain time. The process is derived and evaluated mathematically. The method is verified through numerical tests. Various characteristics are identified to match the dynamic and the static behaviors. For example, the opposite direction of a dynamic load should be considered due to the vibration response. A dynamic bad is transformed to multiple equivalent static loads according to the number of the critical times. The places of the equivalent static load can be different from those of the dynamic load. An optimization method is defined to use the equivalent static loads. The developed optimization process has the same effect as the dynamic optimization which uses the dynamic loads directly. Standard examples are solved and the results are discussed

• Journal of Institute of Control, Robotics and Systems
• /
• v.4 no.2
• /
• pp.235-245
• /
• 1998

This paper presents the fuzzy optimization model based scheduling methodology for the efficient production control of a FFS(FIexible Flow Shop) under the uncertain production environment. To develop the methodology, a fuzzy optimization technique is introduced in which the uncertain production capacity caused by the random events like the machine breakdowns or the absence of workers is modeled by fuzzy number. Since the problem is NP hard, the goal of this study is to obtain the near optimal but practical schedule in an efficient way. Thus, Lagrangian relaxation method is used to decompose the problem into a set of subproblems which are easier to solve than the original one. Also, to construct the feasible schedule, a heuristic algorithm was proposed. To evaluate the performance of the proposed method, computational experiments, based on the real factory data, are performed. Then, the results are compared with those of the other methods, the deterministic one and the existing one used in the factory, in the various performance indices. The comparison results demonstrate that the proposed method is more effective than the other methods.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.18 no.1
• /
• pp.31-40
• /
• 2023

Recent advances in flash technologies, such as 3D processing and multileveling schemes, have successfully increased the flash capacity. Unfortunately, these technology advances significantly degrade flash's reliability due to a smaller cell geometry and a finer-grained cell state control. In this paper, we propose an asymmetric BER-aware reliability optimization technique (aBARO), new flash optimization that improves the flash reliability. To this end, we first reveal that bit errors of 3D NAND flash memory are highly skewed among flash cell states. The proposed aBARO exploits the unique per-state error model in flash cell states by selecting the most error-prone flash states and by forming narrow threshold voltage distributions (for the selected states only). Furthermore, aBARO is applied only when the program time (tPROG) gets shorter when a flash cell becomes aging, thereby keeping the program latency of storage systems unchanged. Our experimental results with real 3D MLC and TLC flash devices show that aBARO can effectively improve flash reliability by mitigating a significant number of bit errors. In addition, aBARO can also reduce the read latency by 40%, on average, by suppressing the read retries.

• Journal of the Korea Convergence Society
• /
• v.12 no.11
• /
• pp.81-90
• /
• 2021

There are various optimization problems in real world and research continues to solve them. An optimization problem is the problem of finding a combination of parameters that maximizes or minimizes the objective function. Harmony search is a population-based metaheuristic algorithm for solving optimization problems and it is designed to mimic the improvisation of jazz music. Harmony search has been actively applied to optimization problems in various fields such as civil engineering, computer science, energy, medical science, and water quality engineering. Harmony search has a simple working principle and it has the advantage of finding good solutions quickly in constrained optimization problems. Especially there are various application cases showing high accuracy with a low number of iterations by improving the solution through the empirical derivative. In this paper, we explain working principle of Harmony search and classify the leading research in recent 3 years, review them according to category, and suggest future research directions. The research is divided into review by field, algorithmic analysis and theory, and application to real world problems. Application to real world problems is classified according to the purpose of optimization and whether or not they are hybridized with other metaheuristic algorithms.