ROUTING AND PACKAGING OPTIMISATION OF PRINTED PRODUCTS USING MACHINE LEARNING ALGORITHMS
Дата
Назва журналу
Номер ISSN
Назва тому
Видавець
КОМП’ЮТЕРНІ ТЕХНОЛОГІЇ ДРУКАРСТВА
Анотація
The study presents an integrated approach to optimising internal production and
logistics processes in printing enterprises, focusing on the combination of routing and
packaging tasks for printed products. The relevance of this research is driven by the
increasing order volumes, the growing complexity of technological workflows, and the
need for enterprises to adapt to the dynamic conditions of digital transformation. In
traditional production models, these processes are considered in isolation, leading to
inefficient resource utilisation, accumulated delays, and reduced productivity. A concept
of comprehensive optimisation is proposed, which emphasises the interconnection
between routing and packaging as two elements of a unified production flow, where
changes in the parameters of one stage inevitably affect the efficiency of the other.
The study focuses on the application of machine learning algorithms clustering,
regression models, reinforcement learning, and evolutionary methods to develop an
adaptive system for managing production flows. The findings indicate that employing
ML technologies allows for the consideration of a wide range of parameters, the
prediction of delays, optimisation of equipment workloads, and the formation of
homogeneous packaging batches in line with actual technological conditions. In
addition, the theoretical principles of routing and packaging in the printing industry are
presented, emphasising the complexity and multifactorial nature of these processes, as
well as their crucial role in maintaining production rhythm and efficiency.
The proposed integrated model operates as a closed-loop feedback system between
the routing and packaging modules, enabling rapid adaptation to changes in production
status, the emergence of urgent orders, and variability in technological operations. The
results indicate that the application of machine learning methods in the printing sector
provides a foundation for enhancing productivity, reducing order lead times, and
minimising logistical costs. The study offers recommendations for the development of
modern intelligent systems for managing printing production, tailored to the demands
of a high-tech and dynamic printed products market.
Опис
In the contemporary context of digital transformation, printing
enterprises face the need for comprehensive optimisation of internal production and
logistics processes, particularly routing and packaging of printed products. In scholarly
research, these processes are traditionally considered separately, mostly within the
scope of heuristic approaches and mathematical models aimed at minimising order
throughput time and ensuring efficient resource utilisation. At the same time, the
increasing variety of products, the growing complexity of technological workflows, and
heightened demands for rapid order processing underscore the need for more flexible,
adaptive solutions capable of handling large volumes of heterogeneous data.
An integrated consideration of routing and packaging as interconnected stages of a
unified production flow is of particular importance. The trajectory of an order between
printing and post-printing stations directly influences the sequence and workload of
packaging operations, while the chosen packaging method, batch formation, and
logistical requirements may, in turn, necessitate adjustments to routing. This
interdependence highlights the need to move from local optimisation of individual
sections towards joint optimisation of routing and packaging decisions at the systemwide level.
In related fields – logistics, warehouse automation, and robotic processing systems
– the effectiveness of comprehensive optimisation has already been confirmed through
widespread implementation of machine learning algorithms, clustering, evolutionary
approaches, and reinforcement learning. In the printing industry, such approaches
remain relatively novel; however, the results obtained demonstrate their considerable
potential in tasks such as operation time prediction, delay probability assessment,
formation of homogeneous product batches, and optimisation of routing. The
application of ML technologies enables the consideration of a large number of
parameters, adaptation to changes in equipment workload, and rapid response to the
emergence of urgent orders.
The development of integrated models combining routing and packaging
optimisation based on machine learning (ML) algorithms and evolutionary methods is
highly relevant. Such a model is expected to construct and adjust routes taking into
account predicted delays, generate optimal packaging solutions using neural networks
and genetic algorithms, and implement a closed-loop feedback system between the
routing and packaging modules. The proposed approach aims to enhance equipment
utilisation, reduce order fulfilment times, and minimise logistical costs, collectively
increasing the competitiveness of printing enterprises.
Тип публікації
Text
Тип текстової публікації
Стаття
ISSN
2411–9210
Ключові слова
Алгоритми. Keywords: machine learning, optimisation algorithms, printing processes, clustering, packaging optimisation, production routing, intelligent management systems.
Бібліографічний опис
Durnyak B., Liakh I., Yavorskyi P., Tsipino A. Routing and packaging optimisation of printed products using machine learning algorithms // КОМП’ЮТЕРНІ ТЕХНОЛОГІЇ ДРУКАРСТВА / COMPUTER TECHNOLOGIES OF PRINTING, 2(54), 2025. с. 54-69