Innovative Issues in Intelligent Systems

The objective of this paper is to give a realistic overview of the current state of the art of intelligent systems in industry based on the experience from applying these systems in a large global corporation. It includes a short analysis of the differences between academic and industrial research, examples of the key implementation areas of intelligent systems in manufacturing and business, a discussion about the main factors for success and failure of industrial intelligent systems, and an estimate of the projected industrial needs that may drive future applications of intelligent systems.

With recent advances in Ambient Intelligence (AmI), it is becoming possible to provide support to a human in an AmI environment. This paper presents an Adaptive Neuro-Fuzzy Inference System (ANFIS) model based scheme, named as prediction of the attention area using ANFIS (PAA_ANFIS), which predicts the human attention area on visual display with ordinary web camera. The PAA_ANFIS model was designed using trial and error based on various experiments in simulated gaming environment. This study was conducted to illustrate that ANFIS is effective with hybrid learning, for the prediction of eye-gaze area in the environment. PAA_ANFIS results show that ANFIS has been successfully implemented for predicting within different learning context scenarios in a simulated environment. The performance of the PAA_ANFIS model was evaluated using standard error measurements techniques. The Matlab^® simulation results indicate that the performance of the ANFIS approach is valuable, accurate and easy to implement. The PAA_ANFIS results are based on analysis of different model settings in our environment. To further validate the PAA_ANFIS, forecasting results are then compared with linear regression. The comparative results show the superiority and higher accuracy achieved by applying the ANFIS, which is equipped with the capability of generating linear relationship and the fuzzy inference system in input-output data. However, it should be noted that an increase in the number of membership functions (MF) will increase the system response time.

In the presented investigation some problems of Intelligent Building Blocks (BBs) integration are considered with emphasis on the Hybrid Intelligent Control System (HICS) design. The main target is to guaranty relevant operability of the complex system into each possible operation situation in autonomous mode. A design method is adopted using Case-Base Reasoning (CBR) approach. Each case has by—tuple description, involving arbitrary situation and corresponding control system design (strategy, structure, building blocks, algorithms and tuning parameters). The design procedure aims to incorporate all available information for the current situation—internal state (environment and equipment conditions), economic circumstances (throughput, quality, objective function), limitations (resources, ecological), environmental interconnections (disturbances, variability). A reconfiguration approach is accepted using situation-based control combined with a sequential improvement. The properties of IBB integration are studied giving prominence to knowledge oriented of them—ontology, case- and rule-based reasoning, intelligent agents and multi-agent systems (MAS). The loose integration is accepted in this investigation. A number of simulation results are presented and discussed for variety of Hybrid Intelligent Control Systems (HICS), which confirm the effectiveness of the proposed approach for complex, non-linear, faulty systems with large variability of feasible situations.

The phenomena of congestion and packet-drops in high-speed communications and computer networks do affect the quality-of-service and overall performance more than ever existing time-delays with uncertain variations. Their control and, possibly, prevention are subject to extensive research ever since the internet is available. Because of the uncertainties and time-varying phenomena, obtaining the accurate and complete information on the network traffic patterns, especially for the multi-bottleneck case, is rather difficult hence learning intelligent controls are needed. One such control is a multi-agent flow controller (MAFC) based on Q-learning algorithm in conjunction with the theory of Nash equilibrium of opponents’ strategies. The other is a model-independent Q-learning control (MIQL) scheme having focus on the flow with higher priority, which also does not need prior-knowledge on communication traffic and congestion. The competition of communication flows with different priorities is considered as a two-player non-cooperative game. The Nash Q-learning algorithm control obtains the Nash Q-values through trial-and-error and interaction with the network environment so as to improve its behaviour policy. The MAFC can learn to take the best actions in order to regulate source flows that guarantee high throughput and low packet-loss ratio. The MIQL control, through a specific learning processing, does achieve the optimum sending rate for the sources with lower priority while observing the sources with higher priority. Designed intelligent controls achieve superior performances in controlling the flows in high-speed networks in comparison to the standard ones and avoid communications congestion.

Imaginary propositional logic (i-logic) is being introduced through which the classical propositional logic (called in the present work real or r-logic) is extended to complex, summary (s-logic), in which the two logics above are interpreted. For this purpose a constraint (axiom) is added to the structures of the two logics—r and i, which connects heir variables and states. The s-logic received provides a possibility all logic equations to be solved which cannot be done in the frame of the classical propositional logic. It is proved that the s-logic has six states, non-equipotent between each other, i.e. it is a multi-valued logic. All possible truth tables for conjunction, disjunction, and implication between the six states and variables of the three logics—r, i, and s are received by the formalisms being introduced. A number of new results are discussed characteristic for the implication in the s-logic. On the base of the truth tables and the rules Modus Ponens and Modus Tollens a number of relations are received for the logical inference in the s-logic which is different in some aspect from those in the r-logic. Examples are given for s-logic application: solving problems in logical equations and the mental activity.

Short remarks on Generalized net theory are given. Some possible applications of the generalized net apparatus as means for modelling of data-mining processes are discussed.

Prism has been developed as a modular classification rule generator following the separate and conquer approach since 1987 due to the replicated sub-tree problem occurring in Top-Down Induction of Decision Trees (TDIDT). A series of experiments have been done to compare the performance between Prism and TDIDT which proved that Prism may generally provide a similar level of accuracy as TDIDT but with fewer rules and fewer terms per rule. In addition, Prism is generally more tolerant to noise with consistently better accuracy than TDIDT. However, the authors have identified through some experiments that Prism may also give rule sets which tend to underfit training sets in some cases. This paper introduces a new modular classification rule generator, which follows the separate and conquer approach, in order to avoid the problems which arise with Prism. In this paper, the authors review the Prism method and its advantages compared with TDIDT as well as its disadvantages that are overcome by a new method using Information Entropy Based Rule Generation (IEBRG). The authors also set up an experimental study on the performance of the new method in classification accuracy and computational efficiency. The method is also evaluated comparatively with Prism.

The topic of the presented research is contemporary threats leading to serious problems in the nearest future. Advantages and disadvantages of typical synthetic advanced analytics methods are investigated, and obstacles are revealed to their distribution in IT, especially in security systems (SS). Security education issues also have been discussed because the high-quality learning inevitably leads to independent research. Original results for juxtaposing statistical versus logical intelligent information processing methods aiming at possible evolutionary fusions are described, and recommendations are made on how to build more effective applications of classical and/or elaborated novel methods: Kaleidoscope, Funnel, Puzzle, and Contradiction. Characteristic peculiarities, advantages, and problems with coordination and control of the investigated group of methods are demonstrated. It is shown that their combination makes possible not only data driven applications, but more complex knowledge guided control of evolutionary computing using nonstandard sets of constraints. The focus is mainly on advanced analytics applications named Puzzle methods and their interactions with other described methods. It is studied aiming at collaborative statistical and logical research based on quantitative method applications, deep processing and effective management of accumulated knowledge. It is shown that applications of intelligent technologies advance the efficiency of statistical applications by using original set of evolutionary methods for data and knowledge fusion. It is shown that all the demonstrated advantages may be successfully combined with well-known methods from big data, advanced analytics, knowledge discovery, data/web/deep data mining or other modern fields. Also, it is shown how the considered applications enhance the quality of statistical inference, reveal the reasons of its effective use, improve the human-machine interaction between the user and system and hence serve the process of gradual but a sustainable improvement of the results. The usage of ontologies is investigated with the purpose of information transfer by a sense in security multiagent environments or to reduce the computational complexity of practical applications. Applications from many fields using the same set of methods have been displayed aiming to show the strength of the domain independent part of the research.

A review of two groups of basic modal type operators, defined over the intuitionistic fuzzy sets, is given. Two new modal operators are introduced for the first time, and some of their properties are discussed. Some open problems are formulated.

The problem of robust output feedback control for a class of uncertain switched fuzzy time-delay systems via common Lyapunov function and multiple Lyapunov function methods is solved. Based on employing Parallel Distributed Compensation strategy, the fuzzy output feedback controllers are designed such that the corresponding closed-loop system possesses stability and robustness for all admissible uncertainties. An illustrative example and the respective simulation results are given to demonstrate the effectiveness and feasible control performance of the proposed design synthesis.