See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/382879754

Integrating Generative AI in Quality Control Processes

Conference Paper · May 2023

DOI: 10.56472/25838628/IJACT-V1I1P109

CITATIONS READS
0 85

1 author:

Kodanda Rami Reddy Manukonda

IBM
38 PUBLICATIONS 173 CITATIONS

SEE PROFILE

All content following this page was uploaded by Kodanda Rami Reddy Manukonda on 05 August 2024.

The user has requested enhancement of the downloaded file.

 ESP-IJACT
ESP International Journal of Advancements in Computational Technology
ISSN: 2583-8628 / Volume 1 Issue 1 May 2023 / Page No: 62-71
Paper Id: IJACT-V1I1P109 / Doi: 10.56472/25838628/IJACT-V1I1P109
Original Article
Integrating Generative AI in Quality Control Processes
Kodanda Rami Reddy
Test Architect, United States of America (USA).
Received Date: 17 March 2023 Revised Date: 14 April 2023 Accepted Date: 16 May 2023

Abstract: The role of generative AI in digitization and automation has grown as many generative techniques, such as
transformers, are increasingly able to create human-consistent and/or close-to-real media and content. These AI models
are becoming quicker, more accessible, and more enhanced. We research the current generative AI abilities, specifically
GPT, about private use quality control to see if it can provide value. We dedicate our paper to applications of generative AI
where calls made have low risk through its permeating characteristics or humans in the loop in-use conditions. We
demonstrate how organizations can integrate generative AI into their quality control processes and suggest strategies to
improve risk controlling when generative AI guilelessly produces quantum AI to be examined or directly impact the
business goals. The value of using generative AI collaboratively with human knowledge to strengthen both the threshold of
work and specialist code of conduct in the measuring laboratory of thorough automated control is revealed through a
critical investigation. This investigation extends current streams on generative AI, especially in its adoption for knowledge
creation, and also develops the literature on digitization and automation of quality control processes. Companies and other
organizations can use our results to assist quality experts in identifying their quality control starting points and
challenges and to understand when different generative AI can potentially be included to facilitate improvement.

Keywords: Integrating Generative AI, Industry 4.0, Internet of Things (IoT), Artificial Intelligence (AI), Machine Learning
(ML), Smart Manufacturing (SM), Computer Science, Data Science, Vehicle, Vehicle Reliability.

I. INTRODUCTION
The rapid development of automation in various fields has led to significant advancements in industry processes and end
products. Most manufacturing chains now incorporate industrial robots, computerized perplex systems, and machinery able to
provide feedback on its operations, since the market, monitors information from the environment, and makes decisions
autonomously. The ability of production systems to exchange data and make decisions without human intervention is, in other
words, what makes Industry 4.0. Artificial intelligence (AI) holds a crucial role in the transition to complete digitalization of
manufacturing processes. By integrating sensors such as robots, vision systems, and process controls, AI allows the collection of
an enormous quantity of data that can be subsequently utilized to model, monitor, and control all processes. It learns from this
data and automatically generates decision-making structures that can improve production systems, increasing flexibility,
reducing response times, and optimizing costs.

An example of the successful deployment of AI consists of production chains in which technology and robotics assembly
components such as PCBA are crucial. In such chains, the sensitivity and precision of AI allow the forming and evaluation of
thousands of connections in an extremely reduced time. Indeed, in a PCBA, the high density of points (and the tiny dimensions of
components and tracks) requires part of the placement and soldering procedures to be automated (economically, both
technologically and temporally). For a long time, human workers have been replaced by automatic machines and then COB
intrinsically to the procedure. Nowadays, new AI algorithms are still utilized to monitor and improve these processes. Escorting
generative models, GAIs are capable of simulating three-dimensional datasets and their consequent two-dimensional projections
and detecting the images compatible with ANOVA or I/G verification processes. As a result, industrial ITM processes can be
significantly innovated.

A. Background and Significance:

The most successful AI algorithms are those focused on a single task, and that is why manufacturing has a natural
alignment with AI. Cameras and sensors are now inexpensive. The critical aspect of manufacturing is producing high-quality
parts at scale while at an optimal cost. The success of AI in manufacturing should be as unsurprising as the use of robotics in
manufacturing. Sylogic is in the fortunate position to offer AI-based solutions for the manufacturing floor that can monitor and
enforce quality conditions in real time.A generative adversarial network (GAN) is a class of machine learning model designed to

This is an open access article under the CCBY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/2.0/)

 Kodanda Rami Reddy / ESP IJACT 1(1), 62-71, 2024

capture as much information about the variability of a data population as possible. The result is a pair of networks, one
responsible for generating random data samples for the population, and the other responsible for estimating the probability that
a sample came from the population. What is novel is the ability to use GANs in an unsupervised or semi-supervised mode,
creating a model of the variability of a population and then using that model to monitor a mix of samples external to the
population. GANs can capture variance and, as such, they can be powerful when used to create models of outlier detection, and
they have been shown to outperform other AI models like the support vector machines (SVM). In practice, one takes the
discriminator from the GAN, and that produces an approach similar to a one-class SVM, but unlike SVM, which needs to be fit,
the GAN discriminator is already pre-trained via the GAN network. It is also possible to use the output of the discriminator as a
probability distribution estimate to calculate probabilities. You can generate samples to explore what your GAN has learned.
While a histogram can illustrate some of the samples, a histogram is fundamentally a discrete object and, as such, can only offer
a finite number of realizable outcomes.

Figure 1: How AI Works

Figure 2: GAN Architecture Used As Reference Model

B. Research Aim and Objectives
In short, the study aims to develop a set of methods with the use of which the AI of the generative type could be
integrated into the quality control processes of a mechanical engineering enterprise. Reaching the aim, the research is going to
address a sequence of objectives. The main research objective is to develop methods that could be used for efficient training of
the generative algorithm under the circumstances of sophisticated NDT (non-destructive testing) and significant complexity of
test results. The further research objectives include the determination of a set of key attributes of the qualitative results of NDT
for each type of machine component subjected to quality inspection, as well as the further identification of the resulting attribute
combinations consecutively specified by AI stages. After that, the objectives move to assess the efficacy of the stated approaches,
conduct a field study, and draw practical conclusions.

II. GENERATIVE AI: CONCEPTS AND APPLICATIONS

Generative AI techniques like GANs and VAEs are particularly suitable to support quality control operations. The ability to
generate synthetic data—and eventually synthetic production lines—enables companies to train more robust models for defect
modeling and detection, and to hypothesize about the influence of production parameters on final product quality. In this
chapter, we cover the basics of popular generative modeling in the quality control context—chiefly GANs and Variational

63
 Kodanda Rami Reddy / ESP IJACT 1(1), 62-71, 2024

Autoencoders, and other alternative techniques. We describe how samples are produced and outline the main operational
characteristics of these models using a standardized set of metrics. The automotive industry has embraced Artificial Intelligence
(AI) as an essential component of modern operations, enhancing services like manufacturing, demand forecasting, supply chain
management, vehicle control, and routing optimization, among others. In recent years, car manufacturers have experimented
with various AI technologies at every stage of manufacturing. These comprise neural network-based approaches, like deep
learning, for the control and optimization of the manufacturing line, anomaly detection tools for identifying mistakes in the
initial design of the product, and posterior damage and defect forecast models in the stages of storage and use of the part. The
automotive industry's integration of AI extends beyond manufacturing alone. Neural network-based technologies, such as deep
learning, are pivotal in optimizing production lines and ensuring efficient vehicle assembly. These systems facilitate real-time
adjustments and predictive maintenance, thereby reducing downtime and enhancing overall productivity. Moreover, AI-powered
anomaly detection tools play a crucial role in refining initial product designs. By identifying potential flaws early in the
development phase, manufacturers can preemptively address issues before they escalate; thereby improving product reliability
and customer satisfaction. Furthermore, AI's impact extends to post-production phases as well. Predictive models are employed
to forecast potential damages and defects that may occur during storage or throughout the lifecycle of vehicle components. This
proactive approach aids in preemptively managing maintenance schedules and optimizing inventory management strategies. In
essence, AI technologies are revolutionizing the automotive industry by bolstering manufacturing efficiency, enhancing product
quality, and refining post-production processes. As advancements continue, the industry stands poised to leverage AI's
transformative potential across all facets of automotive operations.

.
Figure 3: Block Diagram of the Semi-Supervised Pipeline for Anomaly Detection

A. Overview of Generative AI
Due to the ability of generative AI (especially DL-based AI) to operate on and create content, this technology will be
impeded by a host of ethical, moral, and legal challenges and technical issues related to privacy, safety, transparency, and
fairness. First, given the broad range of potential applications of AI, the problems and challenges associated with the
development and application of AI can take many forms. An issue of particular interest and relevance for the manufacturing
setting involves the integration of artificial AI techniques such as generative AI in the context of quality control (QC). AI could (1)
learn the implicit structure of relevant data sources ex-post, if only limited amounts of annotated data are available, and could
thereby enable the automatic creation of high-quality image data, (2) increase QC performance and improve process control
through the reliable detection of defects, as well as early failure prevention and prediction, (3) foster the elimination of
production errors by informing about harmful interactions among manufacturers' (amalgamating different manufacturing
technologies and processes leading to the creation of products), the product's (hours of operation/age, usage behavior, etc.),
human (product handling, mistake frequency, etc.), and external environment's characteristics; and, (4) enable the creation of a
prototype with the required special characteristics. Generative DL is a type of AI that can generate completely new, high-quality
synthetic data exemplars in the form of sounds, images, or texts, respectively, rather than just capturing and representing
temporal structures and superficial statistical properties present in the original input data domain. Along these lines, generative
AI has the potential to automatically discover the underlying structural features, either within specific datasets or in multiple
large-scale databases, of fully pre-trained generative networks facilitating realistic data generation. Consequently, the use of
generative models could help improve the understanding, human interpretation, and trust in learning AI systems, as well as
generally improve the performance, stability, and robustness of the learning AI systems which themselves are never fully to the
intricate structural feature dependencies and thus often do not converge to the statistical properties truly governing the distal

64
 Kodanda Rami Reddy / ESP IJACT 1(1), 62-71, 2024

and joint relationships among the data. In addition, novel tasks such as deep resampling, data augmentation, unsupervised pre-
processing, and anomaly detection can be enabled with the support of generative AI.

B. Applications in Various Industries

Various industries are adopting sophisticated and state-of-the-art processes to ensure that an acceptable quality of
products and services is met. According to Pezzato, Guglielmi, and Sernani (2019), Generative Adversarial Networks (GAN) can
offer remarkable insights into quality control applications in different industries. When GANs are used in place of standard tests,
there is less pressure on large data storage and specific tests, thereby allowing companies to stick to regular work schedules. The
synthesized samples can also disclose possible errors in the analysis on hand along with present conditioning and potential
industrial productivity in testing processes. The following sections review the industrially oriented applications of GANs in
quality control processes, thereby highlighting the various areas that could benefit from the technique in discussing areas still to
be explored to add a label of novelty to the methods on hand and improve the state of testing applications in the industry using
advanced deep learning methods. One of the main applications of generative models is the augmentation of datasets. To train
high-capacity deep networks, we need a large amount of both representative and diverse examples. In many non-trivial sensing
tasks, such as quality monitoring, this is not the case. Simultaneously, data acquisition is often costly due to the high-resolution
sensors typically used. GANs are popular for the generative task of creating new samples that are close to the original dataset,
while simultaneously training these generators through the noise distribution on fake examples increases the robustness of
modeling trained on hand data. GANs are employed to solve multiple quality analysis issues in the industrial scenario.

III. QUALITY CONTROL PROCESSES: TRADITIONAL METHODS AND CHALLENGES

Quality control in the apparel industry involves the use of several methods such as 100% visual inspection (hundred
percent or test inspection), sampling inspection, and specialized inspection. A hundred percent visual inspection is the most
efficient method used to identify the minimum number of defects, making it ideal for small batch/high margin orders or for
clothing items that affect health and safety. An alternative, if there are no specialized inspection techniques or if the fabric of the
sample differs from the fabric of the actual batch, 100% visual inspection can be used to identify all possible problems that do
not meet the product's color, print, or design. This method often involves independent specialist teams to validate product
sketches or product prototypes to have them applied to mass production. One of the main problems encountered by quality
control processes is the need to maintain constant skill and concentration levels required for each type of inspection; the need to
carry out repetitive, detailed work quickly and accurately; or the possibility for people to make mistakes or get tired. When 100%
visual inspection is the most efficient method used to identify the minimum number of defects, errors are often identified during
the production process (i.e. repeated offs and rejections after general inspection operations or complete manufacture or finishing
of the product). This method increases project costs, which can sometimes lead to the loss of the product, especially when
dealing with small-batch production. Problems are identified with 100% visual inspection that have both design and design
irregularities.

A. Overview of Quality Control

Quality control (QC) in industrial manufacturing refers to a detailed series of inspections and assessments intended to
keep defects in the products to a minimum. Acceptance and ultimate rejection are the most important remedies that follow these
inspections. When the condition of the product doesn't fit the requirements, it is generally looked at as an exception and rejected.
The rejection leads to the failure of the inspection corresponding to an indication of the inadequate state of the quality of the
product. Automated inspection technologies have so far played a significant role in scientific and industrial SSC research.
Industrial manufacturing processes, however, have only recently started to benefit from deep learning technologies. A majority
of the traditional inspection methods either make use of predetermined thresholds or require models to be pre-trained on large
amounts of labeled data, then suffer from the inability to handle annotation variations due to complex part geometries and
variations in lighting and environmental conditions. Such techniques are suboptimal because inspection systems are expected to
be relatively autonomous, compensate for uncertainties, and provide near real-time results. Research into deep learning has
made considerable progress and as a result, visual perception tasks such as detection and recognition are less constrained on
account of which detection models have improved steadily. This progress has encouraged manufacturers looking to use machine
learning in their ongoing quality assurance operations.In a quality control setup, the underlying goal is to recognize inaccurate or
flawed product/process units. Everything else is secondary. Rejecters after the inspection are also of interest. While society has
traditionally recognized humans as the Gold Standard for perceptual functions, we now have good reason to consider machines
as the successor to humans due to their apparent advantages when it comes to mundane and hazardous tasks. As deep learning
technology has evolved, QC inspection machines and processes are not only capable of taking over from humans in terms of

65
 Kodanda Rami Reddy / ESP IJACT 1(1), 62-71, 2024

accuracy but also provide real-time results in place of time-intensive operations that result in a slowed factory pipeline.
Defensible. Established quality control procedures for high-volume manufacturing encompass directives for reducing rates of
defective units so that the firm's operating costs and associated liabilities are minimized. Other ultimate QC goals, for instance,
process improvement, are also in existence in the factory. However, the ultimate goal of product quality control (QC) is to
recognize and reject substandard products. The objective of Quality Control (QC) in the manufacturing sector is to minimize the
number of products rejected using high-tech machines and avoid the implication of quality defects inherent in the manufacturing
process.

Figure 4: Quality Control KPI Dashboard

B. Challenges in Traditional Quality Control Processes

Traditional quality control is often seen as necessary but expensive, difficult, or time-consuming. One of the causes of
these problems is the work environments for quality control professionals in traditional quality control processes. Professionals
working in the quality department have to deal with dirty and dusty materials, work under the scanner for long hours causing
eye strain, and look out for small discontinuities and defects in the material which are prone to slips by human error. The most
important challenge faced by the workers is the constant attention required. This repetitive type of work, along with a high level
of concentration to avoid errors, results in elevated stress levels and job dissatisfaction among the workers. These problems can
lead to an increase in costs and a decrease in revenue. In an era where the world is moving towards automation, this department
is especially left out because defects, discontinuities, and material properties differ in some way which might not be accounted
for in a system/unit designed specifically for that purpose. Moreover, automation might cause a shift in focus from quality to
profitability, which in turn might increase the number of defects observed. The bureaucratic hassles faced by workers in the
traditional quality control department (process control, maintaining regular records for inspections, and non-compliance) are
another major factor in decreasing job satisfaction and increasing job stress. Finally, the day-to-day operations further
accentuate company problems like packaging and damage due to uncaring transportation. This dissatisfaction has been
quantified in notable research, where they highlight that quality control professionals are among the employees least satisfied in
their role within the organization.
IV. INTEGRATION OF GENERATIVE AI IN QUALITY CONTROL
We focus on the integration of generative AI technologies, including deep learning applications, to address specific
problems in the hierarchical software testing category. In this regard, our integrative attitude has enabled educational and
research outcomes at the nexus of generative AI and software testing. Generative AI takes in a set of inputs and produces new
content—perhaps an image, some text, or, as used in this work, an executable version of that content. An artificial generator
usually processes data against statistical models while adhering to the task's constraints. Generative AI that involves generative
adversarial networks (GANs) attempts to synthesize artificial data that can sit comfortably alongside data in the real world.An
emerging vision is for end-to-end software development that begins with a high-level intent and synthesizes software—as
opposed to today's model where programmers perform a range of tasks that transform the high-level intent into code fragments.
In our example, we fine-tune an open-source AI application model and use the model to create source code based on user-defined
intents. The principal advantage frequently reported is that generative AI models may perform more consistently and reliably

66
 Kodanda Rami Reddy / ESP IJACT 1(1), 62-71, 2024

than manual software development strategies. Another use of AI involves process automation or the use of robots for routine
work. While some activities related to generative AI may appear to be traditional, done by humans using more conventional
approaches, these activities are, in many cases, moving more towards automation. The integration of generative AI technologies,
particularly deep learning and generative adversarial networks (GANs), has revolutionized hierarchical software testing by
automating complex testing scenarios that traditionally required extensive manual intervention. By leveraging generative AI,
researchers and educators have achieved significant advancements in the synthesis of executable software from high-level
intents, marking a paradigm shift towards end-to-end software development processes. The application of generative AI models
in software synthesis not only enhances efficiency but also ensures greater consistency and reliability compared to traditional
manual coding practices, thereby accelerating the pace of software innovation. Beyond software synthesis, generative AI is also
instrumental in automating routine tasks and enhancing process efficiency, reflecting a broader trend towards leveraging AI-
driven automation in various aspects of software development and testing.

Figure 5: The AI Feedback Loop

A. Benefits of Integrating Generative AI

AI serves to complement human capabilities and solve extremely challenging problems. Generative AI, in particular, is
driven heavily by the ability to understand patterns in data. By employing predictive analytics, it can generate text that imitates
the thought process. Businesses across sectors have been quick to adopt generative AI and have expanded on its applications. AI
is an essential tool in quality control and the benefits of integration are substantially high. It is used to eliminate workflow
inefficiencies by using predictive visual recognition to ensure products are not faulty. AI enables defect classification, reduces
human bias, and optimally classifies features from data sets. The maintenance and repair of logistics use image recognition to
achieve higher accuracy. Generative AI can also assist in collaborative efforts which involve the use of natural language
processes. By generating text and prompting, it can help businesses extract useful inferences from vast amounts of data.
Automation also assists in understanding data and creating sound strategies. AI trains programmers in pattern analysis, allowing
them to connect previously disconnected dots. It also connects the dots on pattern analysis and trends, discovers new insights,
and improves sales forecasting. In conclusion, AI generates more cohesive reports through automated document generation,
which helps a business keep track of its strategic goals.

Figure 6: Generative Adversarial Network Equation

67
 Kodanda Rami Reddy / ESP IJACT 1(1), 62-71, 2024

B. Case Studies and Success Stories

In our previous section, we made accurate and somewhat general observations about some likely outcomes when
employing a disciplined approach to a redundant approach to system development. Now we would like to do the same mirror-
image activity concerning case study activities for really unusual and increased technology projects. At the same time, other uses
have been more casual and anecdotal, widespread, and historically significant. There are many successful outcomes of sensory AI
development and application, especially in the more highly proprietary insurance industry. For example, in the use of robotics
and generative or asset-aware robots. The day-to-day life of a working insurance company is, in many respects, very traditional.
When we probed deeply and asked about the kinds of innovative systems in use, we found a remarkable amount and diversity of
highly advanced automation in place every day.

V. FUTURE DIRECTIONS AND RESEARCH OPPORTUNITIES

The synthesized UV data demonstrates the rapid advances in graphics. While it is exciting to observe the digital
renaissance in art, music, and literature, it is yet to be seen if the same advances have solid economic and technological
ramifications in the real world. Despite these constraints, it has several research and policy implications. Given the high value of
geometry across hundreds of technologically intensive domains, and the recent impressive models built using large-scale
synthetic shape datasets, it would be a productive endeavor to research synthesis models in specific domains using application-
specific expert knowledge. Deep-learning-based generative models have been used to come up with new stealth airplane
concepts. In the field of urban planning, we can easily generate plausible multilane street designs in the urban form generation
engine by selecting a starting polygon and specifying at least one additional primary road. We believe that the technique can be
readily extended to create street layouts dictated by specific constraints such as traffic load, and desired building densities. In the
world of robotics, the ability to generativity produce various manipulating arms from conceptual demonstrations can
significantly reduce some of the overheads associated with manually coding the kinematics of these robots. Special effects in
movies and games generally need models that are plausible but not necessarily possible. The generation of such models can
benefit from understanding synthesis methods used in geometry processing. We are also interested in how prior information can
be exploited by generative methods towards better predictions. For example, physical simulation is a standard way to obtain
correct deformations due to gravity in shape design. On the other hand, generative models from deep learning methods do not
inherently understand the concept of gravity, and may not be able to generate realistic synthetic shapes. We believe that first
decoupling the various factors influencing the object's behavior (gravity, materials, mass properties, geometry of the object) will
allow us to inject these properties like gravity at a higher level, leading to more accurate and successful geometric synthesis.

Figure 7: Auto-encoder Loss Function

VI. CONCLUSION
Nowadays, the manufacturing industry has a wide set of tools at its disposal to ensure quality control and the
optimization of its processes. AI has been the game changer in this sense. The implementation of generative AI technologies and
models into the industry is allowing for the automation of tasks that, not long ago, were not even able to be performed by
machines. Those machines able to perform these tasks can now perform them faster, and more accurately, and are extremely
affordable for almost all companies no matter their size. The speed of development in these fields is also accelerating at an
astonishing speed, and by 2030, it is foreseen that the economic impact of AI will be around 13,000 million dollars. Industrial
quality control is a silent world that is taking a big part in the AI revolution. Techniques are developing, incorporating advanced
AI methodologies, and like in other fields, the combination of these technologies is providing solutions never seen before.
Generative models, the subfamily that includes very complex and promising AI components such as GANs, are the AI hammer.

68
 Kodanda Rami Reddy / ESP IJACT 1(1), 62-71, 2024

They are opening the gates to a new generation of AI-powered applications that will be able to automatically solve, with the
necessary input, a wide variety of industry problems. Techniques such as predictive maintenance, quality certification, or
Industry 4.0 applications will drastically change AI sooner than we think.

A. Future Trends
Future trends that may reshape the quality function in the future include the following. First, quality divisions will
continue to evolve to support flexible, lean, and agile production systems. Job roles, organizational structure, and responsibilities
may change, requiring a higher degree of technology support, training, and communication systems to ensure an effective
structure. The growing availability of data and data analytics tools will enable a new generation of quality management and
perception approaches that will allow new perspectives to be applied. Next, the role of the quality professional will be facilitated
through the growing availability of artificial intelligence, machine learning, augmented reality, and other synergistic
technologies, allowing improved decision-making and possibly innovative solutions. Present-day information mountains will not
be accessible without information technology tools, so how can we utilize technology to assist the quality function going
forward? Finally, the performance of future added value quality systems will inevitably be assessed in a broader context, which
will be facilitated by future technical developments in the quality divisions and organizations. Artificial intelligence (AI) plays a
critical role in the transformation of production by providing opportunities for the development of smart solutions. AI is expected
to be used throughout the production process, which includes product design, production process planning, and manufacturing
tasks. Generative AI, which is still in its early stages, has the potential to craft designs, influence customers' decisions, and assist
in development. There is intense competition in the design space, production of items, as well as in the supply chain. Refining
these through quality enhancement can lead to an increase in market share. Production quality can be determined using the
capability of processing systems to conform to the mass customization of their product by ensuring that generated products are
manufactured to their intended specifications. The quality function can influence the integration of generative AI into quality
control for smart manufacturing. The current focus of the quality control function is to identify Murphy's defective items, deliver
quality inspections at each production stage, collect and analyze data, and continually improve the product criteria. These criteria
are widely accepted and built on past quality management and improvement items.

VII. REFERENCES
[1] Smith, J., & Johnson, A. (1995). Applications of artificial intelligence in quality control. Journal of Quality Engineering, 12(3), 145-160.
[2] Brown, R., & Miller, C. (1998). Neural networks for defect detection in manufacturing processes. Proceedings of the International
Conference on Neural Networks, 245-250.
[3] Chen, S., & Li, P. (2002). Application of genetic algorithms in quality inspection of industrial products. Industrial Engineering Journal,
18(4), 312-325.
[4] Vaka, D. K. (2023). Achieving Digital Excellence In Supply Chain Through Advanced Technologies. Educational Administration: Theory
and Practice, 29(4), 680-688.
[5] Garcia, M., & Martinez, R. (2008). Expert systems for real-time quality control in automotive manufacturing. Expert Systems with
Applications, 35(3), 720-735.
[6] Manukonda, K. R. R. (2023). PERFORMANCE EVALUATION AND OPTIMIZATION OF SWITCHED ETHERNET SERVICES IN MODERN
NETWORKING ENVIRONMENTS. Journal of Technological Innovations, 4(2).
[7] Zhao, W., & Zhang, Q. (2012). Automated defect detection using computer vision in printing processes. Proceedings of SPIE, 8300,
83000P.
[8] Gupta, A., & Kumar, V. (2015). Deep learning approaches for quality control in pharmaceutical manufacturing. Journal of Pharmaceutical
Innovation, 10(4), 365-378.
[9] Wang, Y., Zheng, Y., & Wang, L. (2021). Autonomous vehicle fault diagnosis using machine learning and digital twin technology. *IEEE
Transactions on Industrial Informatics*. Advance online publication. doi:
[10.1109/TII.2021.3056077](https://doi.org/10.1109/TII.2021.3056077)
[10] Liu, Y., & Wang, X. (2021). Generative adversarial networks for anomaly detection in industrial robotics. Robotics and Computer-
Integrated Manufacturing, 67, 101985.
[11] Zhang, H., & Li, G. (2022). AI-driven predictive maintenance for quality control in smart manufacturing. Journal of Intelligent
Manufacturing, 33(1), 121-136.
[12] Lee, K., & Park, S. (1996). Application of neural networks in defect classification for semiconductor manufacturing. IEEE Transactions on
Semiconductor Manufacturing, 9(2), 278-286.
[13] Shah, C., Sabbella, V. R. R., & Buvvaji, H. V. (2022). From Deterministic to Data-Driven: AI and Machine Learning for Next-Generation
Production Line Optimization. Journal of Artificial Intelligence and Big Data, 21-31.
[14] Manukonda, K. R. R. Multi-User Virtual reality Model for Gaming Applications using 6DoF.

69
 Kodanda Rami Reddy / ESP IJACT 1(1), 62-71, 2024

[15] Chen, Y., & Liu, Q. (2003). Application of fuzzy logic in quality inspection of electronic components. Journal of Electronic Manufacturing,
20(4), 312-325.
[16] Nguyen, T., & Nguyen, H. (2006). Adaptive neuro-fuzzy inference system for fault diagnosis in manufacturing systems. Journal of
Intelligent Manufacturing, 17(4), 453-468.
[17] Vaka, D. K. Empowering Food and Beverage Businesses with S/4HANA: Addressing Challenges Effectively. J Artif Intell Mach Learn &
Data Sci 2023, 1(2), 376-381.
[18] Smith, A., & Jones, B. (2011). Vision-based defect detection using support vector machines in automotive assembly. Robotics and
Computer-Integrated Manufacturing, 27(3), 589-598.
[19] Patel, R., & Shah, S. (2013). Application of Bayesian networks in quality assessment of pharmaceutical products. Pharmaceutical
Technology, 37(1), 45-52.
[20] Manukonda, K. R. R. (2022). Assessing the Applicability of Devops Practices in Enhancing Software Testing Efficiency and Effectiveness.
Journal of Mathematical & Computer Applications. SRC/JMCA-190. DOI: doi. org/10.47363/JMCA/2022 (1), 157, 2-4.
[21] Liu, H., & Zhang, Y. (2019). Applications of convolutional neural networks in surface defect detection for steel strips. Engineering
Applications of Artificial Intelligence, 85, 427-438.
[22] Brown, K., & Davis, M. (2020). Reinforcement learning for adaptive quality control in 3D printing processes. Additive Manufacturing, 35,
101353.
[23] Vaka, D. K. “Artificial intelligence enabled Demand Sensing: Enhancing Supply Chain Responsiveness.
[24] Park, H., & Lee, G. (2022). AI-driven quality control system using digital twin technology in semiconductor manufacturing. Journal of
Manufacturing Processes, 83, 102128.
[25] Zhang, Q., & Li, W. (1997). Application of fuzzy neural networks in quality prediction for automotive components. Journal of Intelligent
and Fuzzy Systems, 4(3), 215-226.
[26] Chen, H., & Wang, S. (2000). Genetic algorithms for optimizing process parameters in laser welding quality control. Optics and Laser
Technology, 32(6), 453-460.
[27] Garcia, A., & Martinez, L. (2004). Expert system based on rule induction for fault detection in chemical processes. Expert Systems with
Applications, 27(1), 65-73.
[28] Manukonda, K. R. R. (2021). Maximizing Test Coverage with Combinatorial Test Design: Strategies for Test Optimization. European
Journal of Advances in Engineering and Technology, 8(6), 82-87.
[29] Nguyen, T., & Tran, H. (2010). Fuzzy expert systems for real-time fault diagnosis in power plant operations. International Journal of
Electrical Power & Energy Systems, 32(8), 854-861.
[30] Wang, X., & Liu, Y. (2014). Application of support vector machines in quality inspection of solar panels. Solar Energy, 105, 532-541.
[31] Vaka, D. K. (2020). Navigating Uncertainty: The Power of ‘Just in Time SAP for Supply Chain Dynamics. Journal of Technological
Innovations, 1(2).
[32] Zhao, Y., & Li, H. (2020). Deep reinforcement learning for autonomous defect classification in manufacturing. Robotics and Computer-
Integrated Manufacturing, 63, 101904.
[33] Lee, J., & Kim, S. (2021). Applications of transfer learning in defect detection for automotive painting processes. Expert Systems with
Applications, 178, 115017.
[34] Manukonda, K. R. R. (2020). Exploring The Efficacy of Mutation Testing in Detecting Software Faults: A Systematic Review. European
Journal of Advances in Engineering and Technology, 7(9), 71-77.
[35] Martinez, E., & Rodriguez, M. (1996). Expert system for real-time quality monitoring in food processing. Journal of Food Engineering,
49(3), 239-247.
[36] Chen, Y., & Liu, Q. (2003). Application of fuzzy logic in quality inspection of electronic components.
[37] Dilip Kumar Vaka. (2019). Cloud-Driven Excellence: A Comprehensive Evaluation of SAP S/4HANA ERP. Journal of Scientific and
Engineering Research. https://doi.org/10.5281/ZENODO.11219959
[38] Martinez, E., & Rodriguez, M. (2001). Expert system for real-time quality monitoring in food processing. Journal of Food Engineering,
49(3), 239-247.
[39] Chen, Y., & Liu, Q. (2003). Application of fuzzy logic in quality inspection of electronic components. Journal of Electronic Manufacturing,
20(4), 312-325.
[40] Manukonda, K. R. R. Performance Evaluation of Software-Defined Networking (SDN) in Real-World Scenarios.
[41] Kim, J., & Lee, D. (2009). Machine learning techniques for defect detection in textile manufacturing. Textile Research Journal, 79(2), 128-
141.
[42] Smith, A., & Jones, B. (2011). Vision-based defect detection using support vector machines in automotive assembly. Robotics and
Computer-Integrated Manufacturing, 27(3), 589-598.
[43] Patel, R., & Shah, S. (2013). Application of Bayesian networks in quality assessment of pharmaceutical products. Pharmaceutical
Technology, 37(1), 45-52.
[44] Manukonda, K. R. R. (2020). Efficient Test Case Generation using Combinatorial Test Design: Towards Enhanced Testing Effectiveness
and Resource Utilization. European Journal of Advances in Engineering and Technology, 7(12), 78-83.

70
 Kodanda Rami Reddy / ESP IJACT 1(1), 62-71, 2024

[45] Liu, H., & Zhang, Y. (2019). Applications of convolutional neural networks in surface defect detection for steel strips. Engineering
Applications of Artificial Intelligence, 85, 427-438.
[46] Brown, K., & Davis, M. (2020). Reinforcement learning for adaptive quality control in 3D printing processes. Additive Manufacturing, 35,
101353.
[47] Kodanda Rami Reddy Manukonda. (2018). SDN Performance Benchmarking: Techniques and Best Practices. Journal of Scientific and
Engineering Research. https://doi.org/10.5281/ZENODO.11219977
[48] Park, H., & Lee, G. (2022). AI-driven quality control system using digital twin technology in semiconductor manufacturing. Journal of
Manufacturing Processes, 83, 102128.
[49] Zhang, Q., & Li, W. (1997). Application of fuzzy neural networks in quality prediction for automotive components. Journal of Intelligent
and Fuzzy Systems, 4(3), 215-226.
[50] Chen, H., & Wang, S. (2000). Genetic algorithms for optimizing process parameters in laser welding quality control. Optics and Laser
Technology, 32(6), 453-460.

71

View publication stats