Multimodal Objective Research Articles

적응성 있는 차분 진화에 의한 함수최적화와 이벤트 클러스터링

Differential evolution(DE) has been preyed to be an efficient method for optimizing real-valued multi-modal objective functions. DE`s main assets are its conceptual simplicity and ease of use. However, the convergence properties are deeply dependent on the control parameters of DE. This paper proposes an adaptive differential evolution(ADE) method which combines with a variant of DE and an adaptive mechanism of the control parameters. ADE contributes to the robustness and the easy use of the DE without deteriorating the convergence. 12 optimization problems is considered to test ADE. As an application of ADE the paper presents a supervised clustering method for predicting events, what is called, an evolutionary event clustering(EEC). EEC is tested for 4 cases used widely for the validation of data modeling.

Robust sensor fault diagnosis and validation in HVAC systems

A robust fault diagnosis and validation strategy for temperature sensors and ‘ow meters in central chilling plant is developed, which is based on the first law of thermodynamics. The strategy evaluates soft sensor faults (biases) by examining and minimizing the sum of the squares of concerned mass or steady state energy balance residuals represented by the corrected measurement over a period. It considers systematically all the concerned energy balances and obtains the best estimates of the sensor biases by minimizing the sum of the mean squares of normalized residuals of all energy balances involved. The genetic algorithm technique is employed to determine the global minimal solution to the multimodal objective function, which can be difficult to achieve by traditional gradient-directed search methods. Performance of an advanced robust fault detection, diagnosis and evaluation (FDD&E) scheme is compared with that of a sequential scheme, which was reported earlier, in simulation tests. The robust scheme is superior to the sequential scheme in robustness to abrupt sensor faults, such as biases, etc. The robust scheme is applied to a central chilling plant in an existing commercial building, providing satisfied bias estimates. As a basic method, the sensor FDD&E strategy is of practical value in heating, ventilation and air-conditioning (HVAC) systems as well as in systems where the measurements of liquid ‘ow variables are essential to control and performance monitoring.

Imagery in multi-modal object learning

Spatial objects may not only be perceived visually but also by touch. We report recent experiments investigating to what extent prior object knowledge acquired in either the haptic or visual sensory modality transfers to a subsequent visual learning task. Results indicate that even mental object representations learnt in one sensory modality may attain a multi-modal quality. These findings seem incompatible with picture-based reasoning schemas but leave open the possibility of modality-specific reasoning mechanisms.

Automatic sensor evaluation in BMS commissioning of building refrigeration systems

A strategy and software is developed to automatically diagnose and evaluate the Building Management Systems (BMS) sensors of building refrigeration systems during commissioning or periodical check (recommissioning). The strategy is based on the first law of thermodynamics (i.e., heat and mass balance of water networks). The strategy evaluates soft sensor faults (biases) by examining and minimizing the weighted sum of the squares of the concerned mass and/or steady state energy balance residuals represented by the corrected measurements over a period, on the basis of the measurements downloaded from BMS. A Genetic Algorithm is employed to determine the global minimal solution to the multimodal objective function, which can be difficult to achieve by traditional gradient-directed search methods. The sensor bias estimates, the confidence intervals of bias estimates and the comparisons of the balance residuals before and after the correction are generated by the software to provide a convenient and reliable means for the engineers to check and diagnose the measurement devices of BMS. The strategy, the software configuration and examples of application are presented in this paper.

Automatic Commissioning and Re-Commissioning of BMS Measurement Instruments of Chilling Systems

A strategy and software is developed to automatically diagnose and evaluate the BMS sensors (measurement Instruments) of building central chilling plants systems during commissioning or periodical check (re-commissioning). The strategy is based on the heat and mass balance of water networks (i.e. the first law of thermodynamics), that ensure the robustness of robustness Of the strategy. The strategy evaluates soft sensor faults (biases) by examining and minimizing the weighted sum of the squares of the concerned mass and/or steady state energy balance residuals represented by the corrected measurements over a period, on the basis of the measurements downloaded from BMS. A Genetic Algorithm is employed to determine the global minimal solution to the multimodal objective function. The sensor bias estimates, the Confidence intervals of bias estimates and the comparisons of the balance residuals before and after the correction are generated by the software to provide a convenient and reliable means for the en...

Self-adaptive mutations may lead to premature convergence

Self-adaptive mutations are known to endow evolutionary algorithms (EA) with the ability of locating local optima quickly and accurately, whereas it was unknown whether these local optima are finally global optima provided that the EA runs long enough. In order to answer this question, it is assumed that the (1+1)-EA with self-adaptation is located in the vicinity P of a local solution with objective function value /spl epsi/. In order to exhibit convergence to the global optimum with probability one, the EA must generate an offspring that is an element of the lower level set S containing all solutions (including a global one) with objective function value less than /spl epsi/. In case of multimodal objective functions, these sets P and S are generally not adjacent, i.e., min{/spl par/x-y/spl par/:x/spl isin/P, y/spl isin/S}>0, so that the EA has to surmount the barrier of solutions with objective function values larger than /spl epsi/ by a lucky mutation. It will be proven that the probability of this event is less than one even under an infinite time horizon. This result implies that the EA can get stuck at a nonglobal optimum with positive probability. Some ideas of how to avoid this problem are discussed as well.

Involuntary orienting to sound improves visual perception.

To perceive real-world objects and events, we need to integrate several stimulus features belonging to different sensory modalities. Although the neural mechanisms and behavioural consequences of intersensory integration have been extensively studied, the processes that enable us to pay attention to multimodal objects are still poorly understood. An important question is whether a stimulus in one sensory modality automatically attracts attention to spatially coincident stimuli that appear subsequently in other modalities, thereby enhancing their perceptual salience. The occurrence of an irrelevant sound does facilitate motor responses to a subsequent light appearing nearby. However, because participants in previous studies made speeded responses rather than psychophysical judgements, it remains unclear whether involuntary auditory attention actually affects the perceptibility of visual stimuli as opposed to postperceptual decision and response processes. Here we provide psychophysical evidence that a sudden sound improves the detectability of a subsequent flash appearing at the same location. These data show that the involuntary orienting of attention to sound enhances early perceptual processing of visual stimuli.

The model multiplicity problem: experimenting with real-time specification methods

The object-process methodology (OPM) specifies both graphically and textually the system's static-structural and behavioral-procedural aspects through a single unifying model. This model singularity is contrasted with the multimodel approach applied by existing object oriented system analysis methods. These methods usually employ at least three distinct models for specifying various system aspects: mainly structure, function, and behavior. Object modeling technique (OMT), the main ancestor of the unified modeling language (UML), extended with timed statecharts, represents a family of such multimodal object oriented methods. Two major open questions related to model multiplicity vs. model singularity have been: 1) whether or not a single model, rather than a combination of several models, enables the synthesis of a better system specification; and 2) which of the two alternative approaches yields a specification that is easier to comprehend. The authors address these questions through a double-blind controlled experiment. To obtain conclusive results, real time systems, which exhibit a more complex dynamic behavior than nonreal time systems were selected as the focus of the experiment. We establish empirically that a single model methodology, OPM, is more effective than a multimodel one, OMT, in terms of synthesis. We pinpoint specific issues in which significant diiferences between the two methodologies were found. The specification comprehension results show that there were significant differences between the two methods in specific issues.

Auditory-visual integration during multimodal object recognition in humans: a behavioral and electrophysiological study.

The aim of this study was (1) to provide behavioral evidence for multimodal feature integration in an object recognition task in humans and (2) to characterize the processing stages and the neural structures where multisensory interactions take place. Event-related potentials (ERPs) were recorded from 30 scalp electrodes while subjects performed a forced-choice reaction-time categorization task: At each trial, the subjects had to indicate which of two objects was presented by pressing one of two keys. The two objects were defined by auditory features alone, visual features alone, or the combination of auditory and visual features. Subjects were more accurate and rapid at identifying multimodal than unimodal objects. Spatiotemporal analysis of ERPs and scalp current densities revealed several auditory-visual interaction components temporally, spatially, and functionally distinct before 200 msec poststimulus. The effects observed were (1) in visual areas, new neural activities (as early as 40 msec poststimulus) and modulation (amplitude decrease) of the N185 wave to unimodal visual stimulus, (2) in the auditory cortex, modulation (amplitude increase) of subcomponents of the unimodal auditory N1 wave around 90 to 110 msec, and (3) new neural activity over the right fronto-temporal area (140 to 165 msec). Furthermore, when the subjects were separated into two groups according to their dominant modality to perform the task in unimodal conditions (shortest reaction time criteria), the integration effects were found to be similar for the two groups over the nonspecific fronto-temporal areas, but they clearly differed in the sensory-specific cortices, affecting predominantly the sensory areas of the nondominant modality. Taken together, the results indicate that multisensory integration is mediated by flexible, highly adaptive physiological processes that can take place very early in the sensory processing chain and operate in both sensory-specific and nonspecific cortical structures in different ways.

A global search procedure for parameter estimation in neural spatial interaction modelling

Parameter estimation is one of the central issues in neural spatial interaction modelling. Current practice is dominated by gradient based local minimization techniques. They find local minima efficiently and work best in unimodal minimization problems, but can get trapped in multimodal problems. Global search procedures provide an alternative optimization scheme that allows to escape from local minima. Differential evolution has been recently introduced as an efficient direct search method for optimizing real-valued multi-modal objective functions (Storn and Price 1997). The method is conceptually simple and attractive, but little is known about its behavior in real world applications. This article explores this method as an alternative to current practice for solving the parameter estimation task, and attempts to assess its robustness, measured in terms of in-sample and out-of-sample performance. A benchmark comparison against backpropagation of conjugate gradients is based on Austrian interregional telecommunication traffic data.

Benchmark problems for optimal shape design for 2D eddy currents

Two new benchmark inverse problems for eddy currents are proposed. The first originates in the optimal design of the tubular inductive heater. The authors’ goal is to offer a simple problem which will check whether new software is able to minimize a multimodal objective function. The second benchmark is a simplified model of a hardening device. The purpose of this problem is to test the ability of the software to deal with different design parameters. This benchmark can be easily extended for more complicated, coupled fields problems. For both benchmarks reference standard solutions are presented. They were obtained using a finite element package for electromagnetics, developed by the authors.

A genetic algorithm for multivariate isotonic regression

Isotonic regression problems under partial order restrictions are usually formulated as a minimizing, problem with a multimodal objective function subject to partial order restrictions. Some numerical methods to this kind of problems have been reported when some additional conditions are imposed. However, the classic methods (for example, pool adjacent violators algorithm) are usually not suited to multivariate isotonic regression problems. In order to obtain the global optimal parameters of multivariate isotonic regression, this paper presents a genetic algorithm and illustrates its effectiveness by some numerical examples.

MARS - A MULTISTART ADAPTIVE RANDOM SEARCH METHOD FOR GLOBAL CONSTRAINED OPTIMIZATION IN ENGINEERING APPLICATIONS

A multistart, step-controlled random search algorithm for global, constrained optimization is proposed. The method is found to be very efficient for solving a variety of constrained nonlinear optimization problems. The performance of the method and its comparison with another stochastic search algorithm, simulating annealing, are demonstrated through a number of standard test problems involving multimodal objective functions with continuous and mixed-discrete variables. The applications of the method to a number of practical engineering optimization cases in the field of turbine design and compact heat exchangers are discussed.

Migration velocity analysis using a genetic algorithm

Migration velocity analysis, a method for determining long wavelength velocity structure, is a critical step in prestack imaging. Solution of this inverse problem is made difficult by a multimodal objective function; a parameter space often vast in extent; and an evaluation procedure for candidate solutions, involving the calculation of depth‐migrated image gathers, that can be prohibitively expensive. Recognizing the global nature of the problem, we employ a genetic algorithm (GA) in the search for the optimum velocity model. In order to describe a model efficiently, regions of smooth variation are identified and sparsely parametrized. Region boundaries are obtained via map migration of events picked on the zero‐offset time section. Within a region, which may contain several reflectors, separate components describe long and short wavelength variations, eliminating from the parameter space, models with large velocity fluctuations. Vital to the success of the method is rapid model evaluation, achieved by generating image gathers only in the neighbourhood of specific reflectors. Probability of a model, which we seek to maximize, is derived from the flatness of imaged events. Except for an initial interpretation of the zero‐offset time section, our method is automatic in that it requires no picking of residual moveout on migrated gathers. Using an example data set from the North Sea, we show that it is feasible to solve for all velocity parameters in the model simultaneously: the method is global in this respect also.

Comparison of a genetic algorithm and simulated annealing in an application to statistical image reconstruction

Genetic algorithms (GAs) are adaptive search techniques designed to find near-optimal solutions of large scale optimization problems with multiple local maxima. Standard versions of the GA are defined for objective functions which depend on a vector of binary variables. The problem of finding the maximum a posteriori (MAP) estimate of a binary image in Bayesian image analysis appears to be well suited to a GA as images have a natural binary representation and the posterior image probability is a multi-modal objective function. We use the numerical optimization problem posed in MAP image estimation as a test-bed on which to compare GAs with simulated annealing (SA), another all-purpose global optimization method. Our conclusions are that the GAs we have applied perform poorly, even after adaptation to this problem. This is somewhat unexpected, given the widespread claims of GAs‘ effectiveness, but it is in keeping with work by Jennison and Sheehan (1995) which suggests that GAs are not adept at handling problems involving a great many variables of roughly equal influence. We reach more positive conclusions concerning the use of the GA‘s crossover operation in recombining near-optimal solutions obtained by other methods. We propose a hybrid algorithm in which crossover is used to combine subsections of image reconstructions obtained using SA and we show that this algorithm is more effective and efficient than SA or a GA individually.

燃料電池最適配置計画問題の遺伝的アルゴリズムによる解法

Technological problems of fuel cell generation systems have been overcome mostly; if fuel cell manufactures succeed in reducing the production cost, a number of fuel cells will be installed in distribution systems. Therefore, it is important to prepare a policy to evaluate the merit and demerit of fuel cell installations beforehand. In our previous study, we have proposed a framework to optimally introduce fuel cells into distribution systems. According to this framework, an optimal deployment planning of fuel cells has been formulated as a mixed integer programming problem and the authors have developed a solution algorithm based on the Branch and Bound method. For large scale systems, however, a combinatorial explosion makes the algorithm unable to search the optimal solution within a reasonable computation time. In this paper, a solution method using the genetic algorithm technique is proposed to find suboptimal solutions for large scale problems. Genetic algorithm is belonging to multi-agent search algorithms and shows a robust characteristic against a multi-modal objective function like this problem. The proposed algorithm exploits two kinds of techniques to improve its performance: The first is the diploid population and the second is the efficient initialization. The former contributes the preservation of genetic variety and the latter accelerates the speed of convergence. The proposed algorithm has been applied to various example systems and compared with a simulated annealing method to evaluate their performances with successful results.

Adaptive simulated annealing for the optimal design of electromagnetic devices

From the application of simulated annealing to the optimal design of electromagnetic devices, it is usually found that its convergence speed is low. Based on the analysis for the iterative procedure of simulated annealing, adaptive simulated annealing is put forward, which has a higher convergence speed. When used to solve the optimization problem that arises from the optimal design of a 6 kV fuse-contactor and involves a multimodal objective function, adaptive simulated annealing can not only obtain the global optimal solution but also save 30 percent of the computation time taken by simulated annealing.

Messy simulated annealing

AbstractThe method of simulated annealing is modified so that the concept of messy chromosomes is applied. Constituent genes of messy chromosomes are specified by their respective names (indices) and values (alleles) simultaneously. Unlike simple chromosomes (binary vectors), messy chromosomes may be either under‐ or overspecified with respect to the problem being solved. The messy simulated annealing algorithm is a very robust and efficient stochastic optimization method which is able to find correct minima of deceptive or highly multimodal objective functions. This is shown by way of a number of simulations.

Tool requirement planning in stochastic job shops: A simulated annealing approach

Tool management in shop floor control has been discussed widely in the recent literature. The need for an effective tool management system has been underscored by the impact of tool availability in flexible manufacturing systems and cellular manufacturing shops. With the understanding that most job shops face a budgetary constraint on the units of tools to be maintain, this paper presents an optimization model based on queueing theory for determining the number of units of tools to be maintained in stochastic job shop production systems. The optimization model seeks to minimize total tool related cost under some managerial and operational constraints. To plan tool requirement based on the optimization model, an improved simulated annealing algorithm is presented. Starting with an arbitrary initial solution, the algorithm searches for a solution among the combinatorial alternatives in a controlled random fashion. Unlike the iterative improvement algorithms, the simulated annealing algorithm is applicable to multimodal objective function due to its capability of probabilistic hill-climbing. To demonstrate the performance of the model and the algorithm, a medium-scale numerical example is also discussed in detail.

A hybrid algorithm for finding a global minimum

One of the most important problems in non-linear programming is to find out the global minimum of a given objective function. In this paper, a new hybrid algorithm which combines the random optimization method of Matyas (1965) and one of the well-known ordinary descent algorithms having an effective convergence property (for example, the Fletcher-Reeves conjugate gradient method, the Davidon-Fletcher-Powell quasi Newton method, etc.) is proposed in order to find out a global minimum in as small a number of steps as possible. Several computational experiments on multimodal objective functions are carried out in order to test the efficiency of the proposed hybrid algorithm. The results obtained imply that the proposed hybrid algorithm is useful for finding out a global minimum in a small number of steps. A theorem that predicts convergence to a global minimum is also given.