Introduction:
Fuzzy Logic (FL) is a logical system resembling human cognition. FL’s methodology imitates the manner in which human decision-making includes all intermediate options between YES and NO digital principles. The conventional logic framework that a machine would comprehend requires correct feedback and generates a TRUE or FALSE definitive output, which is similar to the YES or NO of humans. The word fuzzy applies to items that are ambiguous or not obvious. We sometimes face a condition in the natural world where we can’t decide if the state is true or false, their fuzzy logic offers a very precious reasoning versatility. We should take into account the inaccuracies and complexities of any case in this manner. According to Charles Elkan,
“Fuzzy logic is a generalization of normal logic in which somewhere between 0.0 and 1.0.0, a definition can have a degree of validity. Standard logic only refers to premises that are absolutely true (with a degree of truth of 1.0) or utterly untrue (having a degree of truth 0.0). For thinking about fundamentally abstract terms, such as ‘tallness,’ Fuzzy logic is expected to be used.”[1]
Fuzzy Logic in Machines
The useful features of fuzzy logic, particularly in consumer goods, are not in high-level artificial intelligence, but in lower-level machines. Fuzzy controllers are usually introduced as applications that run on generic microprocessors. A few special-purpose microprocessors have been developed that do fuzzy hardware operations specifically, but some of these use binary digital signals (0 or 1) at the lowest level of hardware.[2]
While the application of fuzzy logic can be implemented on several machines, it however in most cases can’t be called a computer in itself, rather it can be a part of a computer. For example, it is used in the aerospace sector for rocket and satellite altitude monitoring and is used in the automobile system for speed control and traffic control. In natural language analysis and numerous intensive Artificial Intelligence applications, fuzzy logic is used and is commonly used in contemporary control systems, such as computer technology. Fuzzy logic is also used in washing machines and washing machines are certainly not computers. The fuzzy logic measures the portion of soil and dirt, the amount of soap and water to be used, the direction of the spin, etc. The machine rebalances the washing load to ensure optimum spinning. Otherwise, it limits the spinning speed if excess is detected. Even the delivery of laundry loads reduces spinning noise. In order to classify the cloth type and adjust the washing cycle accordingly, Neuro-fuzzy logic requires optical sensors to sense the dirt in water and a fiber sensor. It is also used as anti-lock brakes in the cars of Nissan, in risky situations, they use fuzzy logic to control braking dependent on car speed, inertia, wheel velocity, and acceleration.
Fuzzy Logic in Computer
Modern CPUs have ALUs that are very powerful and complex. Along with ALUs, a control unit (CU) is used in modern CPUs. A digital circuit used to conduct arithmetic and logic operations is an arithmetic logic unit (ALU). It constitutes the basic building block of a computer’s central processing unit (CPU). Simple addition, subtraction, multiplication, division, and logic operations, such as OR and AND, are carried out by the ALU. The memory stores instructions and data for the program. The control unit retrieves memory data or instructions and utilizes the ALU operations to execute those orders using that information. For the Boolean truth value of the method, 1.0 is the absolute truth value and 0.0 is the absolute false value. Yet there is no logic for the real truth and the absolute false value in the fuzzy system. But in fuzzy logic, there is so much of an intermediate value that is partly true and partly false.
The founder of fuzzy logic, Lotfi Zadeh, acknowledged that human decision-making requires a multitude of possibilities between YES and NO, unlike algorithms, such as, for example, Definitely Yes, Certainly Not, Cannot Say, etc. To obtain the definite output, the fuzzy logic operates on the levels of input scenarios. Fuzzy logic replicates the human thought process and imitates how an entity will make choices by encouraging a machine to act less accurately and logically than traditional computers do. Machine output is normally ‘true or false’ or ‘yes/no’, etc, but these specific inputs are not required by fuzzy logic to produce useful outputs. Rather, relying on imprecise inputs (e.g., ‘small’, ‘not so big’, ‘large’, ‘larger’), the performance from a microprocessor controller may be formulated mathematically. Fuzzy logic includes a group of rules, in the type of ‘IF-THEN’ sentences.[3]
The significant difference between probabilistic statistics and fuzzy logic is there is no doubt regarding a person’s age, but instead about the extent to which someone fits the ‘a’ category. Certain words, such as ‘big,” rich,” popular’ or ‘dark,’ are only true to a certain degree when applicable to a single person or circumstance. Fuzzy logic attempts to calculate this degree and to allow certain knowledge to be exploited by computers. The advantages of fuzzy logic and its associated innovations are being studied by many people. For circumstances in which traditional logic technologies are not successful, fuzzy logic may be used. Lotfi Zadeh, however, noted that fuzzy logic would not replace computers or methodologies, but rather support them in situations where traditional methods struggle to efficiently solve a problem. One of the advantages of fuzzy control is that it can be applied to a normal computer quickly.[4] There are several consumer goods that are using fuzzy logic in their service. Several fuzzy logic processors are constructed without using traditional computers to perform special tasks.
Fuzzy logic can be a part of a computer, however electronic appliances using fuzzy logic cannot be classified as a computer. For example, Mistubishi’s microwave and plasma etching use fuzzy logic to set cooking strategies and lunes power and to set strategy and etch time respectively. Machine learning methods are versatile and simple to apply in Fuzzy logic. Fuzzy logic, when treating both computational evidence and linguistic information at the same time, has contributed to beverage research. Fuzzy logic can create models that reflect the interactions between consumer desires and sensory characteristics. It should be remembered that the more general the nature of fuzzy logic rules is, the greater the number of decision parameters. But fuzzy logic operates on reliable as well as imprecise knowledge, so that consistency is undermined much of the time.
Application
In terms of its great effect on network management, few approaches have been suggested to ensure a network lifespan. Besides, because current systems are primarily dependent on the combination of various parameters, these do not offer additional services, like real-time communications and stable energy usage between sensor nodes, so adaptability concerns remain unresolved between nodes in wireless sensor networks. A fuzzy logic framework providing real-time communication in an assured WSN lifetime was proposed to address this problem. The suggested fuzzy logic controller accepts the energy, time, and velocity of the input descriptors to decide the position of each node for the next length and the next-hop relay node for real-time packets. Via the simulation results, it was confirmed that the current fuzzy logic model effectively guarantees both the lifespan of the assured network and real-time distribution. The paradigm with a new secured pre-configured lifelong system with real-time packet support in WSNs shows that fuzzy logic is an efficient and precise framework that can be effectively extended to any WSN emergency app for effective and effective real-time decisions. Fuzzy logic retains a high degree of accuracy and stable energy usage among sensor nodes in a network field compared to other approaches that have crisp values. Besides, the architecture of the fuzzy-logic system is very simple and straightforward, allowing users/applications to specify various variables, sets, and laws, using the rich energy sensor nodes for real-time operation, based on each setting and the sensor functionality, and showing that almost all sensors used their full energy level by achieving a pre-configured lifespan.[5]
In the Japanese industry, appliance control using fuzzy logic inference and a new class of sensors has had a profound influence on producers, retailers, and customers. Companies could fathom creating the opportunity to take advantage of a rapid change in the course of the consumer for integrating these innovations into appliances. Three key concerns must be resolved by a development strategy to ensure that this change of focus is a pleasant one. First, producers of appliances must target appliances with embedded “smart” capabilities for production. Second, sensor manufacturers must be ready to build sensors that are suitable for fuzzy logic control. Third, theoreticians of conventional controls must establish a central understanding of the philosophy of fuzzy logic inference regulation to study where and how it will be best used.
Resource Management
Shifting and reducing energy demand continues to create tremendous interest for end-users at the household level, especially because of the ongoing design of a competitive pricing strategy. In particular, end-users must serve as a starting point to minimize their usage at peak hours in order to eliminate the need to increase the grid and thus save significant costs. The Fuzzy logic algorithm can effectively forecast the use of short-term loads (STLC). This technique is the foundation of a new algorithm for home energy management (HEM) that can optimize the cost of electricity usage while reducing peak demand. The fuzzy logic modeling requires a heavy dependency on a complete database of data from multiple instrumented show homes for actual consumption. A suggested HEM algorithm helps every end-user to control, with a high degree of versatility and clarity, their energy usage and “reshape” the load profile. This can, for instance, be done predominantly by the smart control of a storage device combined with the remote management of electrical appliances. The findings of the simulation show that an accurate STLC forecast offers the probability of achieving optimum HEM system preparation and operation.[6]
A fuzzy logic algorithm can accurately forecast day-to-day energy use in individual homes. The modeling includes a full database of actual energy usage from several display houses fitted with instruments. To maximize the forecast, the concept of fuzzy rules is an essential aspect. A modern approach with a logic table consisting of “IF-THEN” laws, such as “IF temperature is low”, provides the fuzzy logic system, THEN electricity consumption is high. The approach of fuzzy logic seems to be a perfect way to forecast energy usage, since it is possible to perceive human actions as unpredictable and, at the same time, predictable. The scheme of prediction based on the fuzzy logic process consists of the following steps:
- Fuzzification is used to translate digital inputs into fuzzy inputs.
- To get a deeper understanding of the users’ preferences, the fuzzy rules are required to stem from a learning process.
- The fuzzy inference that uses the rules information to find the consumption of electricity.
- To transform the fuzzy values into digital values, defuzzification is used.[7]
Neuro-fuzzy and Chaos Logic
A neuro-fuzzy model is based on a fuzzy system equipped with a learning algorithm developed from the principle of the neural network. In the fundamental fuzzy scheme, the learning process works on local knowledge and induces only local changes. It is possible to view a neuro-fuzzy system as a 3-layer feedforward neural network. Input variables are represented by the first layer, fuzzy laws are represented by the middle (hidden) layer and output variables are represented by the third layer. Fuzzy sets are structured as weights for (fuzzy) relations. It may, however, be easy, since it describes inside the model the data flow of information processing and learning. As a system of fuzzy rules, a neuro-fuzzy system will still be viewed. The framework can also be built from scratch using training data since it can be initialized by prior information in the form of fuzzy rules. The neuro-fuzzy mechanism approximates an undefined dimensional function which the training data partly describes. Unclear samples represent the fuzzy rules embedded inside the system, which can be used as models of the training results. A neuro-fuzzy system cannot be used as a kind of specialist (fuzzy) system because, in the strict sense, this has nothing to do with fuzzy logic.[8]
Neuro-fuzzy hybridization ends in a hybrid artificial system that blends these two approaches with the learning and connectivism framework of neural networks by incorporating the human-like thinking style of fuzzy systems. In the research, neuro-fuzzy hybridization is generally called a fuzzy neural network (FNN) or neuro-fuzzy structure (NFS). E.g., mixing a neural network with a fuzzy system leads to a neuro-hybrid system that results in a neuro-fuzzy hybrid model.
Chaotic systems create large numbers of behavior patterns and are abnormal since these patterns are switched between them. They are sensitive to initial conditions, which in practice means that chaotic systems can switch fairly fast among patterns. Today’s electronic computers carry out computations based on operational activities of digital logic enacted as logic circuits at the lowest point, indicating that almost all calculations can be conducted by sequences of ones and zeros. Inevitably, all computations in computers are conducted in hardware with configurations of these gates that execute logic functions, as easy as adding two figures and as complex as reformatting a file. As logic gates, there are seven simple logic functions put in place. By selecting a single chaotic component, each one of the logical operations is discovered. Through integrating logical gates, computer memory could be built. It is feasible to create a simple, fast, but cost-effective, particular computing system with these fundamental materials in place.
Digital Diary incorporates chaos logic technology while a Simputer applies fuzzy logic and chaos logic together. Other appliances such as a PDA device and a digital camera incorporates Neuro-Fuzzy Logic.
Conclusion
Fuzzy Logic’s key contribution to machine learning is its graduality, which is the ability to reflect gradual principles and characteristics. Fuzzy principles in a specific problem domain are used as modeling components. The granularity provided by fuzzy methods helps to generate information granules for the representation of knowledge in machine learning, another essential property. Fuzzy sets provide usability by offering a connection between numerical values and linguistic words consisting of symbolic values. Inevitably, in several real-world contexts, like medical decision-making, education, multimedia application, etc., fuzzy logic coupled with machine learning can be implemented. Because of these incentives, machine learning based on fuzzy logic is a compelling approach to improved decision making.[9]
References:
[1] n.d, “What is ‘fuzzy logic’? Are there computers that are inherently fuzzy and do not apply the usual binary logic?”, Scientific American (1999), https://www.scientificamerican.com/article/what-is-fuzzy-logic are-t/ (last visited Dec 2020).
[2] Ibid.
[3] P.J. Fellows, Properties of food and principles of processing, Food Processing Technology, 4, 2017.
[4] n.d, Supra
[5] Babar Shah, et al, “Fuzzy logic-based guaranteed lifetime protocol for real-time wireless sensor networks.”, 15.8, Sensors, 20373-20391, (2015), https://www.mdpi.com/1424-8220/15/8/20373/htm
[6] Sébastien Bissey, et al, “The fuzzy logic method to efficiently optimize electricity consumption in individual housing.”, 10.11, Energies, 1701, (2017), https://www.mdpi.com/1996-1073/10/11/1701/pdf
[7] Id.
[8] What are Neuro-Fuzzy Systems?, http://fuzzy.cs.ovgu.de/nfdef.html (last visited Jan 2021)
[9] Sreeja Ashok, et al, “Fuzzy Based Machine Learning: A Promising Approach”, 41.8, CSI COMMUNICATIONS,21,24, (2017).
0 Comments