Cluster Events
NEXT SEMINAR 2024/2025
In this section, you will find the information (Speaker/Guest, abstract, date & time slot , location, online registration link..) about the upcoming seminar during the year. Everyone is welcome, from within ESSEC as well as from outside.
Claudia O. Lopez Soto
Tuesday, January 14th, 2025 from 12:10pm to 1:10 pm
Room : N230 - New signage
Zoom meeting : https://essec.zoom.us/j/99808564251
Title : Beam Search to Minimise Cutting Patterns with Maximum Utilisation
Abstract : In many cutting and packing problems the sole objective is to maximise the utilisation of the cutting board, while satisfying the demand of pieces to be cut. When one focuses on optimising the use of raw material, it is very likely that for each board we will have to enter different cutting coordinates. In operations that involve a large number of pieces, many boards will be required, so the cutting process can be highly affected when changing boards. When dealing with a heterogeneous mix of pieces, this problem cannot be avoided. However, when the set of pieces to be cut presents large repetitions of the same pieces, it would seem reasonable to pay attention to the number of changes in the machine settings, reducing the cutting process time. In this problem, we introduce a new objective in the classic two-dimensional bin packing problem, where on top of maximising the total utilisation, we are also interested in minimising the number of different patterns needed to complete the demand. For this work, a pattern is defined as the coordinates of a set of pieces that will be cut from a bin. Minimising the number of patterns, will result in less changes on the settings of the cutting machine among bins, with a reduction on total cutting times. This variant of the problem has not been widely studied and so far we have only found the work by Song & Bennell (2014) which uses a column generation model to solve this problem. To start tackling this problem, we expand the initial work in Bennell et al. (2018) and work with a similar setting based on the glass industry, with irregular pieces that can be freely rotated and reflected. We also start by considering the restriction of separating each piece by means of a guillotine cut, which reduces the solution space and somehow limits the placement of each piece. Bennell et al. (2018) present a beam search strategy where each node represents a complete bin, thus the branch with minimum depth is the one with fewer bins, or maximum utilization. The same idea applies in this problem, interpreting the branch depth as the number of changes in machine settings. We introduce some novelty approaches in terms of the sorting of pieces, considering not just the largest ones as the first ones to place, but also taking into consideration the number of repetitions of each piece. The fact that pieces are separated by guillotine cuts, translates into irregular shapes where to place pieces after a guillotine cut is performed, thus we also look at the similarity between the angles of the pieces and the angles of each section of the bin to decide which piece to place next. Preliminary tests will be shown to demonstrate the efficiency of these new criteria for sorting pieces. We also provide our results, which confirm that the strategy of dealing with both objectives simultaneously is more effective than using existing heuristics that are mainly designed for heterogeneous mix of pieces.
For any information about OMOR events, please send an e-mail to matta@essec.edu.
***
GEMMA BERENGUER FALGUERA - Universidad Carlos III de Madrid
Wednesday, January 29th 2025 from 12pm to 1pm
Room : G.022 - New signage
Zoom meeting : https://essec.zoom.us/j/99964872070
Title : Diversity-wishing in the supply chain: a cross-cultural study
Abstract : In this talk, I will describe the state of supplier diversity efforts by corporations from an international perspective using two different datasets. First, we examine data for the companies in the 2020 and 2022 Fortune Global 500 and find that corporate definitions of diversity are dynamic and vary across regions. Furthermore, while supplier diversity efforts are not yet widespread, these initiatives are increasingly common, especially in the form of references to diversity in companies’ supplier codes of conduct (SCC). Companies in North America and in certain economic sectors, such as the financial and healthcare sectors, are more likely to have such efforts in place. Based on our data, companies that report on their internal diversity and have other forms of supplier sustainability initiatives are also more likely to have supplier diversity initiatives.
In the second study, we expand our company data to include almost ten thousand organizations. We employ fully automated processes for collecting and analyzing SCCs, utilizing Natural Language Processing (NLP) techniques. We also design an innovative method for tone classification to efficiently and accurately extract, classify, and assess the assertiveness of enforcement language related to the SCCs of these companies. This unique data shows that larger and more profitable organizations are likelier to have detailed and more assertive enforcement diversity language. However, the relationships between regional variations and industry sectors and more assertive language are less consistent in this data.
For any information about OMOR events, please send an e-mail to matta@essec.edu.