Sébastien Thomassey

A hybrid sales forecasting system based on clustering and decision trees

Competition and globalization imply a very accurate production and sourcing management of the Textile-Apparel-Distribution network actors. A sales forecasting system is required to respond to the versatile textile market and the needs of the distributors. Nowadays, due to the specific constraints of the textile sales (numerous and new items, short life time), existing forecasting models are generally unsuitable or unusable. We propose a forecasting system, based on clustering and classification tools, which performs mid-term forecasting. Performances of our models are evaluated using real data from an important French textile distributor.

A neural clustering and classification system for sales forecasting of new apparel items

The Textile-Apparel-Distribution network actors require a very accurate production and sourcing management to minimize their costs and satisfy their customers. For a such strategy, distributors rely on sales forecasting system to respond to the versatile textile market. However, the specific constraints of the textile sales (numerous and new items, short lifetime) complicate the forecasting procedure and distributors prefer to use intuitive estimation methods of the sales rather than the existing forecasting models. We propose a decision aid system, based on neural networks, which automatically performs item sales forecasting. Performances of our model are evaluated using real data from an important French textile distributor.

A short and mean-term automatic forecasting system--application to textile logistics

In order to reduce their stocks and to limit stock out, textile companies require specific and accurate sale forecasting systems. More especially, textile distribution involves different forecast lead times: mean-term (one year) and short-term (one week in average). This paper presents two new complementary forecasting models, appropriate to textile market requirements. The first model (AHFCCX) allows to automatically obtain mean-term forecasting by using fuzzy techniques to quantify influence of explanatory variables. The second one (SAMANFIS), based on a neuro-fuzzy method, performs short-term forecasting by readjusting mean-term model forecasts from load real sales. To evaluate forecasts accuracy, our models and classical ones are compared to 322 real items sales series of an important ready to wear distributor.

A simulation based comparison: Manual and automatic distribution setup in a textile yarn rewinding unit of a yarn dyeing factory

Abstract This paper aims to explain the production flow and the distribution logic of bobbins for rewinding process in a yarn dyeing factory, comparing the different scenarios of production (manual and automatic) using the computer simulation tools. The goal of this project is to build a model in which all the involved processes can be simulated with the consideration of all the parameters and constraints. The simulation model is used as a tool for the comparison of present manual setup and future automated setup for the production management of bobbin distribution in yarn rewinding process in terms of delays and costs. Since, the manual operation involves defaults, improper time management, errors and with the growing competitiveness globally, the companies in Europe need to automate as much as possible their production lines. The expected impacts are to increase the productivity and profitability, to have the possibility to customize the production, to develop production tools, implementation of the lean manufacturing tools.

Introduction: Artificial Intelligence for Fashion Industry in the Big Data Era

With the emergence of the big data era, companies, and more especially fashion companies, are faced with a new relationship between consumers, suppliers, and competitors. Fashion companies have also to manage different data with many and complex correlations and dependencies between them and uncertainties related to human factors. It is crucial for companies to master these data flows to optimize their decision making. In such situations, artificial intelligent techniques are particularly efficient. The potential applications of artificial intelligence in fashion industry cover a wide scope from design support systems to fashion recommendation systems through sensory evaluation, intelligent tracking systems, textile quality control, fashion forecasting, decision making in supply chain management or social networks and fashion e-marketing. Thus, this book aims to illustrate the different possibilities and advantages of artificial intelligence for the fashion industry in the big data era. This introduction chapter provides a brief description of each chapter of this book.

A Discrete Event Simulation Model with Genetic Algorithm Optimisation for Customised Textile Production Scheduling

This chapter aims to explain the methodology of the production schedule optimisation for the automatic manufacturing of customised textile products. The data involved in this manufacturing process are huge and constitute many parameters and constraints. The proposed system could be divided into two main modules, the optimisation model and the production floor model. Indeed, the complexity of this scenario demands a hybrid model which involves a combination of an optimisation model (genetic algorithm model) and a production simulation model (discrete event simulation) with a robust link (ActiveX/OLE Automation Server). The system forms a complex synchronised loop which replicates and improves the production schedule in process till the best results are achieved. The expected impacts are to have on-time shipment, increased productivity and profitability with the implementation of lean tools. Indeed, the implementation of this model is very vast. This would permit the use of a powerful discrete event model with an optimisation algorithm which gives numerous possibilities from manufacturing scheduling to the global supply chain, distribution and logistics planning and optimisation.

AI for Apparel Manufacturing in Big Data Era: A Focus on Cutting and Sewing

In the fashion industry, the apparel manufacturing part contains four main processes involving cutting, sewing, finishing, and packing. The complex system deals with configuration of numerous operations and resources in facing of various uncertainties and under constraints of sequence, quantity, time, and cost. Artificial intelligence (AI) has been applied to provide optimal scenario in shorter time than traditional mathematical methods. Big data is helpful due to the ability of prediction for unraveling uncertainties which ensures a smooth and stable production. For improvement of apparel manufacturing in modern fashion industry, it is necessary to develop the capabilities of advanced computing technologies and take great advantage of valuable information that can be dug out from big data.

3D Adaptive Morphotype Mannequin for Target Population

The aim of this paper is to present a design methodology to obtain parametric geometrical model of 3D virtual mannequin representing the human bodies of a target population. The morphological evolution of these bodies depends on the chosen sizes imposed by the needs of the apparel industry. The population is composed of a representative sample of 5108 women aged 18-65 years from the French measurement campaign in 2003. These women have been measured by a 3D body scanner to detect with a good precision the measurements of the bodies. The proposed method starts by choosing a person representing the morphology nearest to the morphotype of the population. The morphotype is characterized by specific measurements from the standard size of the measurement chart. A geometrical model associated with reverse engineering techniques have been realized to generate the 3D virtual parametric mannequin from the 3D body scanned of the morphotype. Then, the volume of the parametric mannequin is managed by the morphological evolution rules extracted from the measurement chart of the classification. These rules are directly applied on the di erent morphological curves which are either located on the anthropometric points or managed by anthropometric and proportion rules. These two sets of parameters are directly applied on the 3D human mannequin in order to manage his volume according to the di erent sizes with a good precision. Finally, the 3D shape of the mannequin is a surface model hung to a morphological curves network. This process guarantees a maximum e ciency and an exceptional morphological similarity of the morphotype when the parametric mannequin evolves from one size to another size.

Qualitative Analysis of Open- Ended Questions to Define Awareness of Ethical Fashion in Romania

This research focuses on consumer behaviour analysis of Romanians as there is a lack of knowledge about their awareness of ethical fashion. The fashion industry and its concept of fast fashion, have negative impacts on the environment and society. The concept of slow or ethical fashion evolved, but consumer behaviour is important to reduce the impacts. Research in consumer behaviour aims at identifying consumer attitudes and consumption patterns. To understand customers and their expectations are increasingly important for the fashion business. This approach can analyse consumer awareness of ethics in fashion among Romanians froman everyday perspective. The author conducted a cross-sectional study, a semi-structured focus group discussion with twenty participants during an organised workshop at a Romanian university using the city of Iasi exemplarily to represent a greater urban area. The dynamic panel discussions stimulated the data collection, including the author as a moderator directing the conversation. Simultaneously, a survey containing qualitative items, i.e., six openended questions, was carried out. This paper discusses findings from this qualitative analysis, focusing on phrases and words most commonly mentioned by the respondents. Data is based on fourteen participants to gather the perception and knowledge of the ethical fashion concept. Results of this empirical social research show that the responses support the theory of ethical fashion, but awareness is limited, and barriers towards sustainable consumption exist, this aligns with existing literature. This study serves as a preliminary research revealing a variety of conclusions and can be useful for further quantitative survey design.

Environmentally-Friendly Perception of Fashion Products: A Kansei Study

One of the biggest challenges of fashion industry is how to become more sustainable. On the one hand, eco has become a trend. The rising awareness of its impact on earth has triggered the creation of environmentally-friendly fashion. More people show interest into products designed with a lower environmental impact. On the other hand, there is confusion among designers and consumers on sustainable creation and image. Also, online shopping has seen a rise. This study analyzes the perception of consumers towards eco values in fashion designs through a survey using kansei engineering, applied to products from a French webshop for kidswear, reflecting their online communication through product images and descriptive words. To analyze the environmentally-friendly image, ten fashion samples were selected, and three main adjectives were chosen: organic, green, recycled/reduced, and divided into subcategories. Design needs to attract its target consumers. The physical appearance of products, i.e., the form, offers meaning to consumers and impacts their purchase behavior. Fashion designers need to create and communicate this visual information. Thus, the research aims to identify how sustainable design is perceived by consumers in terms of eco image.