Banca de DEFESA: THIAGO DA SILVA PINAFFI

Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.
STUDENT : THIAGO DA SILVA PINAFFI
DATE: 26/06/2026
TIME: 09:00
LOCAL: Bloco 19
TITLE:

Development and application of a metaheuristic algorithm for the 2D cutting-stock problem at a Brazilian small and medium-sized furniture maker

KEY WORDS:

Two-dimensional cutting stock problem; Genetic algorithm; Systematic literature review; Small and medium-sized enterprises; Furniture industry;

PAGES: 175
BIG AREA: Engenharias
AREA: Engenharia de Produção
SUBÁREA: Pesquisa Operacional
SPECIALTY: Programação Linear, Não-Linear, Mista e Dinâmica
SUMMARY:

The cutting stock problem consists of allocating smaller items into larger objects without overlap, while respecting specific constraints and pursuing a defined objective, such as minimizing the number of bins used. Although it has been studied since the 1960s and has significant economic relevance in sectors such as sheet metal and furniture, the literature on the topic remains fragmented in two aspects: the lack of comprehensive systematic reviews and the rarity of real-world applications, especially in Small and Medium-sized Enterprises (SMEs). To address these gaps, this dissertation proposes two complementary articles. The first conducts a systematic literature review on the two-dimensional cutting and packing problem, using the Methodi Ordinatio method to select and rank the most relevant studies, analyzing 116 articles out of 447 classified as eligible. The main findings include: the predominance of exact methods (45 articles) over metaheuristics (23 articles); the lack of real industrial applications (only 14 out of the 116 articles); and the near absence of studies applied to SMEs (only three articles). Based on the review findings, the second article proposes a Genetic Algorithm for the two-dimensional cutting problem with threestage guillotine cuts, an unlimited number of bins, and no item rotation, applied both to instances from the literature and to real instances from a Brazilian SME in the furniture sector. Compared with the literature, the algorithm achieves a global gap of 1.89%, with performance ranging from 0% in Class 9 instances to 6.36% in Class 6 instances. For the real instances, the algorithm obtains results equivalent to human planning in all tested cases, with an average reduction of 94.06% in solution time. The method is executed through a command-line interface with an Excel file as input, dispensing with advanced computational infrastructure or specialized technical knowledge, making it directly applicable to the context of SMEs. Taken together, the two articles demonstrate that the methodological sequence systematic review-development-industrial application constitutes a rigorous and replicable approach for advancing the field of cutting and packing problems.

COMMITTEE MEMBERS:
Interna - 122100 - GISLAINE CAMILA LAPASINI LEAL
Externo ao Programa - 242982 - MAURO HENRIQUE MULATI - nullPresidente - ***.837.099-** - RAFAEL HENRIQUE PALMA LIMA - NÃO INFORMADO
Externo ao Programa - ***.967.539-** - REGINALDO FIDELIS - NÃO INFORMADO