In the present work, the directions for expanding the intellectual capabilities of fuzzy control systems of complex technological objects are shown. Often, the operation of the control object is carried out either in ...
详细信息
In the present work, the directions for expanding the intellectual capabilities of fuzzy control systems of complex technological objects are shown. Often, the operation of the control object is carried out either in the conditions of incomplete information, or significant deficiencies in the mathematical description, both in terms of the system of differential equations and in the restrictions imposed on the system as a whole. Overcoming these shortcomings with the use of typical fuzzy approaches is fraught with difficulties associated both with the algorithmic complexity of the knowledge base, the information redundancy of the linguistic variables used, and a significant number of fuzzy terms in a given control range. In the implementation of generally accepted control procedures based on soft calculations, difficulties arise with the implementation of the adaptive properties of such systems. The combination of various algorithms of fuzzy inference allows solving intellectual control problems for multicriteria and multifactor problems. At the same time, a different combination of such fuzzy conclusions enhances the various properties of such systems, for example, robustness, multitasking, aggregate control in the areas of large and small signals, etc. The paper presents models of two systems with a combination of algorithms for fuzzy logic inference Sugeno-Sugeno and Sugeno-Mamdani, analyzes the main characteristics of control systems using such multi-stage systems, identifies advantages and disadvantages for various combinations of algorithms, presents simulation results. In addition, some difficulties and features of setting up an external cascade, based on Sugeno's fuzzy logic inference algorithm, were identified, which can later be eliminated through the use of alternative output algorithms, for example, Mamdani. Features of the application of an algorithm in the internal cascade are most often dictated by the specifics of the operation of the control object,
The problem of instability areas of the general purpose drive system with PWM inverter and an induction motor was investigated. The first task of detection of instability (or excessive angular velocity oscillations wi...
详细信息
The problem of instability areas of the general purpose drive system with PWM inverter and an induction motor was investigated. The first task of detection of instability (or excessive angular velocity oscillations with a poor damping factor) was solved in the drive system without velocity feedback and without velocity signal. The fuzzy inference based algorithm uses for processing only one signal from power circuit-inverter supply current signal, and two other signals from the control circuit of the inverter. It activates the adequate reference frequency correcting procedure. The improvement of stability is achieved only by the influence on the reference frequency signal.
An increasing number of applications require cooperative agents to reason about the state of an distributed uncertainty domain. However, inference process of such system could become overly slow for practical applicat...
详细信息
An increasing number of applications require cooperative agents to reason about the state of an distributed uncertainty domain. However, inference process of such system could become overly slow for practical applications, and there has been significant interest in developing faster approximation techniques. In this paper, we focus on the existing MSBN models for cooperative reasoning in multi-agent environments. We show that, while the MSBNs provide a framework for exact inference, existing algorithms are usually not feasible in larger problem domains. Therefore, we investigate the issues related to the design of efficient inference algorithm for the MSBN model. We then propose a suite of algorithms for approximate multi-agent probabilistic reasoning in MSBNs. Our approach includes an MSBN subnet calibration process and distributed stochastic sampling on MSBN LJFs.
Many problems exist whose solutions take the form of patterns that may be expressed using grammars (e.g., speech recognition, text processing, genetic sequencing). Construction of these grammars is usually carried out...
详细信息
Many problems exist whose solutions take the form of patterns that may be expressed using grammars (e.g., speech recognition, text processing, genetic sequencing). Construction of these grammars is usually carried out by computer scientists working with domain experts. In the case when there is a lack of domain experts, grammar inference can be applied. In this paper, two grammar inference algorithms are briefly described and their application to software engineering is presented.
The failure of traditional expert system algorithms and mathematics in real-time tasks is described, and an alternative inference engine design more appropriate to real-time problem solving is explored. The approach i...
详细信息
The failure of traditional expert system algorithms and mathematics in real-time tasks is described, and an alternative inference engine design more appropriate to real-time problem solving is explored. The approach is generic, but to facilitate its description an example is used involving human supervisory operators of a semiautomated telecommunication network. Each human's real-time process control tasks are first decomposed into several types of subtasks under cases of certain, uncertain, and conflicting information. Two divergent calculi are seen to be required for these various tasks: 1) the "situational calculus" for real-time deterministic control, and 2) the "calculus of uncertainty" for fusion of conflicting information from diverse knowledge sources and for propagating uncertainty through the inference net to reach conclusions that can be executed by the situational calculus. The distributed architecture ramifications of a dual-calculus appoach are explained, and the fusion technique is elaborated.
VLSI floorplanning design automation based on soft computing is discussed. Authors have proposed the fusion of fuzzy inference and genetic algorithms for the automation of VLSI floorplanning design. In this paper, the...
详细信息
VLSI floorplanning design automation based on soft computing is discussed. Authors have proposed the fusion of fuzzy inference and genetic algorithms for the automation of VLSI floorplanning design. In this paper, the expansion of inference is presented. The fuzzy rules are used to infer the initial position of the on-blocks based on analysis of the accumulated knowledge of the expert design engineer. Only the dominant combinations of place and block are inferred. In the authors' previous work (1998, 1999), blocks deemed suitable candidates for placement at the center, relative to the four corners and side of the chip, are inferred. In addition to those inference, such blocks that are relatively appropriate to be placed along with the four perimeters (edges) of active area of the chip. These inferences are then reflected in the initial population of the genetic algorithms. The rest of the block placement phase is entrusted to the genetic algorithms. Experimental software to implement the proposed approach was developed. The results of the experiments showed a level and quality of placement close to that of the expert design engineer.
This paper introduces new and practically relevant non-Gaussian priors for the Sparse Bayesian Learning (SBL) framework applied to the Multiple Measurement Vector (MMV) problem. We extend the Gaussian Scale Mixture (G...
详细信息
ISBN:
(数字)9798350368741
ISBN:
(纸本)9798350368758
This paper introduces new and practically relevant non-Gaussian priors for the Sparse Bayesian Learning (SBL) framework applied to the Multiple Measurement Vector (MMV) problem. We extend the Gaussian Scale Mixture (GSM) framework to model prior distributions for row vectors, exploring the use of shared and different hyperparameters across different measurements. We propose Expectation Maximization (EM) based algorithms to estimate the parameters of the prior density along with the hyperparameters. To promote sparsity more effectively in a non-Gaussian setting, we show the importance of incorporating learning of the parameters of the mixing density. Such an approach effectively utilizes the common support notion in the MMV problem and promotes sparsity without explicitly imposing a sparsity-promoting prior, indicating the methods’ robustness to model mismatches. Numerical simulations are provided to compare the proposed approaches with the existing SBL algorithm for the MMV problem.
Short video platforms push content to users through recommendation algorithms, which greatly improves user experience, but also triggers the information cocoon effect. This paper explores the impact of the recommendat...
详细信息
ISBN:
(数字)9798331504205
ISBN:
(纸本)9798331504212
Short video platforms push content to users through recommendation algorithms, which greatly improves user experience, but also triggers the information cocoon effect. This paper explores the impact of the recommendation algorithm of short video platforms on the information cocoon effect based on the theory of causal inference, and proposes algorithm design and model simulation methods to decouple the relationship between the information cocoon effect and content diversity. First, this paper constructs a causal inference model to identify the formation mechanism of the information cocoon effect in the recommendation algorithm. Then, by introducing a diversity constraint factor, adjustments are made in the recommendation algorithm to increase the probability that users receive different types of content, thereby reducing the generation of the information cocoon effect. The model simulation results show that without affecting the user experience, the optimized recommendation algorithm improves the information diversity index by 15.7 % , while the intensity of the information cocoon effect is reduced by 22.4%. This study shows that the optimization of the recommendation algorithm based on causal inference can effectively alleviate the information cocoon effect in the short video platform, improve the breadth of user information acquisition, and has high practical value.
Auction mechanism, as a fair and efficient resource allocation method, has been widely used in varieties trading scenarios, such as advertising, crowdsensoring and spectrum. However, in addition to obtaining higher pr...
详细信息
Auction mechanism, as a fair and efficient resource allocation method, has been widely used in varieties trading scenarios, such as advertising, crowdsensoring and spectrum. However, in addition to obtaining higher profits and satisfaction, the privacy concerns have attracted researchers' attention. In this paper, we mainly study the privacy preserving issue in the double auction market against the indirect inference attack. Most of the existing works apply differential privacy theory to defend against the inference attack, but there exists two problems. First', indistinguishability' of differential privacy (DP) cannot prevent the disclosure of continuous valuations in the auction market. Second, the privacy-utility trade-off (PUT) in differential privacy deployment has not been resolved. To this end, we proposed an attack-defense game-based reinforcement learning privacy-preserving method to provide practically privacy protection in double auction. First, the auctioneer acts as defender, adds noise to the bidders' valuations, and then acts as adversary to launch inference attack. After that the auctioneer uses the attack results and auction results as a reference to guide the next deployment. The above process can be regarded as a Markov Decision Process (MDP). The state is the valuations of each bidders under the current steps. The action is the noise added to each bidders. The reward is composed of privacy, utility and training speed, in which attack success rate and social welfare are taken as measures of privacy and utility, a delay penalty term is used to reduce the training time. Utilizing the deep deterministic policy gradient (DDPG) algorithm, we establish an actor-critic network to solve the problem of MDP. Finally, we conducted extensive evaluations to verify the performance of our proposed method. The results show that compared with other existing DP-based double auction privacy preserving mechanisms, our method can achieve better results in both privacy
EdgeAI represents a compelling approach for deploying DNN models at network edge through model partitioning. However, most existing partitioning strategies have primarily concentrated on homogeneous environments, negl...
详细信息
EdgeAI represents a compelling approach for deploying DNN models at network edge through model partitioning. However, most existing partitioning strategies have primarily concentrated on homogeneous environments, neglecting the effect of device placement and their inapplicability to heterogeneous settings. Moreover, these strategies often rely on either data parallelism or model parallelism, each presenting its own limitations, including data synchronization and communication overhead. This paper aims at enhancing inference performance through a pipeline system of devices through leveraging both parallel and sequential relationships among them. Accordingly, the problem of Multi-Device Cooperative DNN inference is formulated by optimizing both device placement and model partitioning, taking into account the unique characteristics of heterogeneous edge resources and DNN models, with the goal of maximizing throughput. To this end, we propose an evolutionary device placement technique to determine the pipeline stage of devices by enhancing a variant of particle swarm optimization. Subsequently, an adaptive model partitioning strategy is developed by combining intra-layer and inter-layer model partitioning based on dynamic programming and the input-output mapping of DNN layers, respectively, to accommodate edge resource limitations. Finally, we construct a simulation model and a prototype, and the extensive results demonstrate that our proposed algorithm outperforms current state-of-the-art algorithms.
暂无评论