Michel Happiette

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.

Analysis of the temporal decomposition procedure for scheduling with release and due dates

This paper first presents an improved method of temporal decomposition for minimising the searching space of scheduling problems. Next, the effects of the temporal decomposition procedure in scheduling problems are analysed. It is theoretically shown that the complexity of a scheduling algorithm using decomposed subsets varies inversely with that of the decomposition procedure. Therefore, the efficiency of the overall scheduling algorithm is strongly related to the decomposability of the set of operations to be processed on each machine. This decomposability is evaluated using a probabilistic approach where the probability distributions of the scheduling parameters are obtained from historical workshop data. The average number of decomposed subsets and the average size of these subsets are estimated. Both theoretical analysis and simulation results have revealed that the decomposition procedure leads to optimal effects when some conditions on scheduling parameters are met.

Modelling and Simulation of the Textile Channel by HyperNets

To achieve better communication between those involved in the various aspects of textile-apparel distribution (TAD), a general methodology for the analysis and for the management simulation of the logistic flows must be implemented by the different participants. The model derived in this paper is based on a high-level Petri-net graphical representation, and several simulations are realised among the TAD participants.

Decomposing a Jobshop into Flowshop Structures Using Genetic Algorithms

Abstract This paper presents a method for decomposing a Jobshop production structure into a number of Flowshop structures so that the optimal layout of production can be found in each Flowshop. This decomposition is considered as a clustering problem, which is solved using a genetic algorithm. It permits to optimize a specified objective function related to the overall number of machines used and the number of additional machines inserted for obtaining admissible solutions of layout of production.

A Structural Framework for the Temporal Decomposition of the Single Machine Scheduling Problem

Abstract This paper deals with the problem of scheduling n jobs on a single machine. The proposed approach is not concerned with the formulation and the solution of any optimization problem, but merely with the characterization of some basic properties of feasible solutions, which can be of some help for the human operators. This leads to determine the set of schedules which are compatible with the earliest starting times and the latest finishing times of the jobs. The problem is N.P. complete in the strong sense. In order to simplify the construction process of admissible sequences, we propose here a formulation of the problem based on a digraph which represents the relation between the jobs and the positions (or ranks). The canonical decomposition of this digraph permits to partition the set of jobs into a number of subsets among which a total order is defined.

Sales Forecasting Under Uncertain Environment Fuzzy Classification in Textile Distribution

Abstract This paper presents a methodology for the management of buffer-stocks for textile items by a sales partition approach. The proposed approach uses a fuzzy clustering method to find the optimal classes, where items belonging to each class have similar sales graph, so as to design the same forecasting method. The compactness of the generated classes are constrained to be lower than a compactness index given by the industrial manager, which characterizes the required classification accuracy. An application for Textile/Gannent/Distribution industry is treated.

Temporal Decomposition by Hierarchical Procedure for Scheduling

Abstract This paper presents an hierarchical procedure for temporal decomposition. This procedure permits to transform high dimensional scheduling problem into low dimensional scheduling sub problems. During the decomposition procedure, we search for the best partition according to the principle of minimizing the cost of loss of free margins and product overflows.