• Title/Summary/Keyword: objective function

Search Result 4,500, Processing Time 0.049 seconds

A Method of Determining the Scale Parameter for Robust Supervised Multilayer Perceptrons

  • Park, Ro-Jin
    • Communications for Statistical Applications and Methods
    • /
    • v.14 no.3
    • /
    • pp.601-608
    • /
    • 2007
  • Lee, et al. (1999) proposed a unique but universal robust objective function replacing the square objective function for the radial basis function network, and demonstrated some advantages. In this article, the robust objective function in Lee, et al. (1999) is adapted for a multilayer perceptron (MLP). The shape of the robust objective function is formed by the scale parameter. Another method of determining a proper value of that parameter is proposed.

Comparison of Cost Function of IMRT Optimization with RTP Research Tool Box (RTB)

  • Ko, Young-Eun;Yi, Byong-Yong;Lee, Sang-Wook;Ahn, Seung-Do;Kim, Jong-Hoon;Park, Eun-Kyung
    • Proceedings of the Korean Society of Medical Physics Conference
    • /
    • 2002.09a
    • /
    • pp.65-67
    • /
    • 2002
  • A PC based software, the RTP Research Tool Box (RTB), was developed for IMRT optimization research. The software was consisted of an image module, a beam registration module, a dose calculation module, a dose optimization module and a dose display module. The modules and the Graphical User Interface (GUI) were designed to easily amendable by negotiating the speed of performing tasks. Each module can be easily replaced to new functions for research purpose. IDL 5.5 (RSI, USA) language was used for this software. Five major modules enable one to perform the research on the dose calculation, on the dose optimization and on the objective function. The comparison of three cost functions, such as the uncomplicated tumor control probability (UTCP), the physical objective function and the pseudo-biological objective function, which was designed in this study, were performed with the RTB. The optimizations were compared to the simulated annealing and the gradient search optimization technique for all of the optimization objective functions. No significant differences were found among the objective functions with the dose gradient search technique. But the DVH analysis showed that the pseudo-biological objective function is superior to the physical objective function when with the simulated annealing for the optimization.

  • PDF

Development of Optimization Methodology for Laser Welding Process Automation Using Neural Network Model and Objective Function (레이저 용접공정의 자동화를 위한 신경망 모델과 목적함수를 이용한 최적화 기법 개발)

  • Park, Young-Whan
    • Transactions of the Korean Society of Machine Tool Engineers
    • /
    • v.15 no.5
    • /
    • pp.123-130
    • /
    • 2006
  • In manufacturing, process automation and parameter optimization are required in order to improve productivity. Especially in welding process, productivity and weldablity should be considered to determine the process parameter. In this paper, optimization methodology was proposed to determine the welding conditions using the objective function in terms of productivity and weldablity. In order to conduct this, welding experiments were carried out. Tensile test was performed to evaluate the weldability. Neural network model to estimate tensile strength using the laser power, welding speed, and wire feed rate was developed. Objective function was defined using the normalized tensile strength which represented the weldablilty and welding speed and wire feed rate which represented the productivity. The optimal welding parameters which maximized the objective function were determined.

POI Recommendation Method Based on Multi-Source Information Fusion Using Deep Learning in Location-Based Social Networks

  • Sun, Liqiang
    • Journal of Information Processing Systems
    • /
    • v.17 no.2
    • /
    • pp.352-368
    • /
    • 2021
  • Sign-in point of interest (POI) are extremely sparse in location-based social networks, hindering recommendation systems from capturing users' deep-level preferences. To solve this problem, we propose a content-aware POI recommendation algorithm based on a convolutional neural network. First, using convolutional neural networks to process comment text information, we model location POI and user latent factors. Subsequently, the objective function is constructed by fusing users' geographical information and obtaining the emotional category information. In addition, the objective function comprises matrix decomposition and maximisation of the probability objective function. Finally, we solve the objective function efficiently. The prediction rate and F1 value on the Instagram-NewYork dataset are 78.32% and 76.37%, respectively, and those on the Instagram-Chicago dataset are 85.16% and 83.29%, respectively. Comparative experiments show that the proposed method can obtain a higher precision rate than several other newer recommended methods.

A novel PSO-based algorithm for structural damage detection using Bayesian multi-sample objective function

  • Chen, Ze-peng;Yu, Ling
    • Structural Engineering and Mechanics
    • /
    • v.63 no.6
    • /
    • pp.825-835
    • /
    • 2017
  • Significant improvements to methodologies on structural damage detection (SDD) have emerged in recent years. However, many methods are related to inversion computation which is prone to be ill-posed or ill-conditioning, leading to low-computing efficiency or inaccurate results. To explore a more accurate solution with satisfactory efficiency, a PSO-INM algorithm, combining particle swarm optimization (PSO) algorithm and an improved Nelder-Mead method (INM), is proposed to solve multi-sample objective function defined based on Bayesian inference in this study. The PSO-based algorithm, as a heuristic algorithm, is reliable to explore solution to SDD problem converted into a constrained optimization problem in mathematics. And the multi-sample objective function provides a stable pattern under different level of noise. Advantages of multi-sample objective function and its superior over traditional objective function are studied. Numerical simulation results of a two-storey frame structure show that the proposed method is sensitive to multi-damage cases. For further confirming accuracy of the proposed method, the ASCE 4-storey benchmark frame structure subjected to single and multiple damage cases is employed. Different kinds of modal identification methods are utilized to extract structural modal data from noise-contaminating acceleration responses. The illustrated results show that the proposed method is efficient to exact locations and extents of induced damages in structures.

Development of Pareto strategy multi-objective function method for the optimum design of ship structures

  • Na, Seung-Soo;Karr, Dale G.
    • International Journal of Naval Architecture and Ocean Engineering
    • /
    • v.8 no.6
    • /
    • pp.602-614
    • /
    • 2016
  • It is necessary to develop an efficient optimization technique to perform optimum designs which have given design spaces, discrete design values and several design goals. As optimization techniques, direct search method and stochastic search method are widely used in designing of ship structures. The merit of the direct search method is to search the optimum points rapidly by considering the search direction, step size and convergence limit. And the merit of the stochastic search method is to obtain the global optimum points well by spreading points randomly entire the design spaces. In this paper, Pareto Strategy (PS) multi-objective function method is developed by considering the search direction based on Pareto optimal points, the step size, the convergence limit and the random number generation. The success points between just before and current Pareto optimal points are considered. PS method can also apply to the single objective function problems, and can consider the discrete design variables such as plate thickness, longitudinal space, web height and web space. The optimum design results are compared with existing Random Search (RS) multi-objective function method and Evolutionary Strategy (ES) multi-objective function method by performing the optimum designs of double bottom structure and double hull tanker which have discrete design values. Its superiority and effectiveness are shown by comparing the optimum results with those of RS method and ES method.

Multi-objective Optimization of Vehicle Routing with Resource Repositioning (자원 재배치를 위한 차량 경로계획의 다목적 최적화)

  • Kang, Jae-Goo;Yim, Dong-Soon
    • Journal of Korean Society of Industrial and Systems Engineering
    • /
    • v.44 no.2
    • /
    • pp.36-42
    • /
    • 2021
  • This paper deals with a vehicle routing problem with resource repositioning (VRPRR) which is a variation of well-known vehicle routing problem with pickup and delivery (VRPPD). VRPRR in which static repositioning of public bikes is a representative case, can be defined as a multi-objective optimization problem aiming at minimizing both transportation cost and the amount of unmet demand. To obtain Pareto sets for the problem, famous multi-objective optimization algorithms such as Strength Pareto Evolutionary Algorithm 2 (SPEA2) can be applied. In addition, a linear combination of two objective functions with weights can be exploited to generate Pareto sets. By varying weight values in the combined single objective function, a set of solutions is created. Experiments accomplished with a standard benchmark problem sets show that Variable Neighborhood Search (VNS) applied to solve a number of single objective function outperforms SPEA2. All generated solutions from SPEA2 are completely dominated by a set of VNS solutions. It seems that local optimization technique inherent in VNS makes it possible to generate near optimal solutions for the single objective function. Also, it shows that trade-off between the number of solutions in Pareto set and the computation time should be considered to obtain good solutions effectively in case of linearly combined single objective function.

Improvement of learning method in pattern classification (패턴분류에서 학습방법 개선)

  • Kim, Myung-Chan;Choi, Chong-Ho
    • Journal of Institute of Control, Robotics and Systems
    • /
    • v.3 no.6
    • /
    • pp.594-601
    • /
    • 1997
  • A new algorithm is proposed for training the multilayer perceptrion(MLP) in pattern classification problems to accelerate the learning speed. It is shown that the sigmoid activation function of the output node can have deterimental effect on the performance of learning. To overcome this detrimental effect and to use the information fully in supervised learning, an objective function for binary modes is proposed. This objective function is composed with two new output activation functions which are selectively used depending on desired values of training patterns. The effect of the objective function is analyzed and a training algorithm is proposed based on this. Its performance is tested in several examples. Simulation results show that the performance of the proposed method is better than that of the conventional error back propagation (EBP) method.

  • PDF

Robust Optimization Using Supremum of the Objective Function for Nonlinear Programming Problems (비선형계획법에서 목적함수의 상한함수를 이용한 강건최적설계)

  • Lee, Se Jung;Park, Gyung Jin
    • Transactions of the Korean Society of Mechanical Engineers A
    • /
    • v.38 no.5
    • /
    • pp.535-543
    • /
    • 2014
  • In the robust optimization field, the robustness of the objective function emphasizes an insensitive design. In general, the robustness of the objective function can be achieved by reducing the change of the objective function with respect to the variation of the design variables and parameters. However, in conventional methods, when an insensitive design is emphasized, the performance of the objective function can be deteriorated. Besides, if the numbers of the design variables are increased, the numerical cost is quite high in robust optimization for nonlinear programming problems. In this research, the robustness index for the objective function and a process of robust optimization are proposed. Moreover, a method using the supremum of linearized functions is also proposed to reduce the computational cost. Mathematical examples are solved for the verification of the proposed method and the results are compared with those from the conventional methods. The proposed approach improves the performance of the objective function and its efficiency.

Full waveform inversion by objective functions with power and integral (지수 및 적분을 포함한 목적함수에 의한 파형역산)

  • Ha, Wan-Soo;Pyun, Suk-Joon;Shin, Chang-Soo
    • 한국지구물리탐사학회:학술대회논문집
    • /
    • 2007.06a
    • /
    • pp.130-134
    • /
    • 2007
  • Classical full waveform inversion for velocity estimation defines the objective function as the $l^2$ -norm of differences between the modeled and the observed wavefields. Although widely used, the results of this method have been less than satisfactory. A moderate improvement of this method is to define the objective function as the $l^2$ -norm of differences between the logarithms of the modeled and observed wavefields. In this paper we propose new objective functions of waveform inversion. They produce better results in sub-salt imaging than those of the classical and the logarithmic objective functions. One objective function defines the residual as the difference between $L^{th}$ power of the modeled wavefields and that of the observed wavefields. Another defines the residual as the difference between the integral of the $L^{th}$ power of the modeled wavefields and that of the observed wavefields. We apply these new objective functions to the synthetic SEG/EAGE salt model, and show that our new waveform inversion algorithms provide more accurate results than those of the classical and logarithmic waveform inversion methods.

  • PDF