Paul Gaborit

Scheduling rules to minimize total tardiness in a parallel machine problem with setup and calendar constraints

Quality control lead times are one of most significant causes of loss of time in the pharmaceutical and cosmetics industries. This is partly due to the organization of laboratories that feature parallel multipurpose machines for chromatographic analyses. The testing process requires long setup times and operators are needed to launch the process. The various controls are non-preemptive and are characterized by a release date, a due date and available routings. These quality processes lead to significant delays, and we therefore evaluate the total tardiness criterion. Previous heuristics were defined for the total tardiness criterion, parallel machines, and setup such as apparent tardiness cost (ATC) and ATC with setups (ATCS). We propose new rules and a simulated annealing procedure in order to minimize total tardiness.

Concurrent product configuration and process planning, towards an approach combining interactivity and optimality

In mass customisation, defining concurrently the configured product and the planning of the associated production process is a key issue in the customer/supplier relationship. Nevertheless, few studies propose supporting the decision-maker during the resolution of this significant problem. After studying the decision-makers needs and problem characterisation (modelling and scale aspects), we propose in this paper a two-step approach with the aid of some tools. The first step allows the customer or internal requirements to be captured interactively with a constraint-based approach. However, this step does not lead to one single solution, e.g. there are many uninstantiated remaining decision variables. In this paper, we suggest adding an original optimisation step to complete this task. Thus, the contribution of the study is twofold: first, methodologically to define a new two-step approach that meets industrial needs; and second, to provide adapted tools especially for the optimisation step. The optimisation step, using a multi-criteria constrained evolutionary algorithm, allows the user to select their own cost/cycle time compromise among a set of Pareto optimised solutions. A conventional evolutionary algorithm is adapted and modified, with the inclusion of filtering processing, in order to avoid generating invalid solutions. Experimentations are described, and a comparison is made with a branch-and-bound approach that outlines the interest in the propositions.

How to take into account general and contextual knowledge for interactive aiding design: Towards the coupling of CSP and CBR approaches

The goal of this paper is to show how it is possible to support design decisions with two different tools relying on two kinds of knowledge: case-based reasoning operating with contextual knowledge embodied in past cases and constraint filtering that operates with general knowledge formalized using constraints. Our goals are, firstly to make an overview of existing works that analyses the various ways to associate these two kinds of aiding tools essentially in a sequential way. Secondly, we propose an approach that allows us to use them simultaneously in order to assist design decisions with these two kinds of knowledge. The paper is organized as follows. In the first section, we define the goal of the paper and recall the background of case-based reasoning and constraint filtering. In the second section, the industrial problem which led us to consider these two kinds of knowledge is presented. In the third section, an overview of the various possibilities of using these two aiding decision tools in a sequential way is drawn up. In the fourth section, we propose an approach that allows us to use both aiding decision tools in a simultaneous and iterative way according to the availability of knowledge. An example dealing with helicopter maintenance illustrates our proposals.

Configuration of high performance apartment buildings renovation: A constraint based approach

This paper looks at the possibility of configuring high performance renovation of apartment buildings by the use of a constraint satisfaction problem (CSP). A new thermal envelope wraps the whole building and allows it to achieve a really low energy performance of 25 kWh/m2/year. This new thermal envelope is composed of rectangular panels, always comprising insulation and cladding, and sometimes including, in addition, doors, windows or solar modules. The panels can be fixed directly onto the walls or onto a metal structure around the whole building. The decision support system, based on CSP approaches, will simultaneously enable the interactive definition of the renovation, the associated bill of material (BOM) and the building site assembly process. The range of knowledge to exploit and to model leads us to integrate into a single decision support system different constraint types, such as classical, dynamic, geometric and global constraints, as well as their filtering methods.

Evaluation and design: a knowledge-based approach

The aim of this communication is to describe how aiding-design tools can evaluate designed solutions to help users make the best choices, avoid design mistakes and reduce the design time-cycle. First, we will compare the two main methods for aiding design—behaviour simulation tools and domain knowledge simulation tools—and look at their advantages and drawbacks. We will focus on tools based on knowledge because of their ‘interactivity’ and for their ability to represent domain knowledge and show how they can be extended to evaluate designed solutions. We will then concentrate on an aiding-design tool based on constraints and see how a solution can be evaluated using an evaluation function. As such a tool has already been developed as part of a European project to help metallurgists design and evaluate heat treatment operations, we end with the presentation of a real example.

A Recursive Algorithm for Building Renovation in Smart Cities

Layout configuration algorithms in civil engineering have two major strategies called constructive and iterative improvement. Both strategies have been successfully applied within different facility scenarios such as room configurations and apartment layouts. Yet, most of the work share two commonalities: They attack problems in which the reference plane is parallel to the Earth and, in most cases, the number of activities are known in advance. This work aims to close that gap by developing a constructive-based algorithm for the layout configuration of building facades in the context of a French project called CRIBA. The project develops a smart-city support system for high-performance renovation of apartment buildings. Algorithm details are explained and one example is presented to illustrate the kind of facades it can deal with.

Aiding design with constraints: an extension of quad trees in order to deal with piecewise functions

This paper deals with aiding preliminary design when considered as a constraint satisfaction problem (CSP). In this case, constraint filtering techniques provide some kind of interactive assistance to the designer. However, some kinds of numerical constraints corresponding with numerical relations cannot be filtered precisely with classical analytical filtering techniques such as interval arithmetic or box-consistency; it is therefore necessary to discretize them in order to include them in the CSP. To this end, quad trees (QT) have been proposed for binary constraints, or 2k trees when more than two variables are considered; but QT assume that a constraint must be defined by a single numerical function. The aim of this paper is to show that QT techniques can be extended when a constraint is defined by a piecewise function or by a set of numerical functions defined on intervals. The first section recalls some basics relevant to the preliminary design problem and the interests of the CSP assistance. The second section presents the principles of the QT. The last section describes our contributions relevant to QT extensions dealing with piecewise functions.

Building renovation adopts mass customization: Configuring insulating envelopes

This work is motivated by an industrial need of manufacturing façades insulating envelopes in order to reduce energy consumption in residential buildings. An insulating envelope is a configuration of a set of rectangular panels that respects a set of limitations. Due to the number of façades to be renovated and the number of possible configurations for a single façade, the envelope configuration is both a mass customization problem as well as a combinatorial one. The paper then introduces a decision support system based on the framework of constraint satisfaction, as it fits neatly the constrained nature of the problem. Two configuration tasks have been identified as prerequisite to envelopes configurations: (1) the configuration of a questionnaire for information inputs and (2) the configuration of a constraint satisfaction problem for each one of the façades to be renovated. The system architecture promotes maintenance, modularity and efficiency as different configuration tasks are divided into web-services. Conception and implementation of the massive building thermal renovation are then supported.

Open Packing for Facade-Layout Synthesis Under a General Purpose Solver

Facade-layout synthesis occurs when renovating buildings to improve their thermal insulation and reduce the impact of heating on the environment. This interesting problem involves to cover a facade with a set of disjoint and configurable insulating panels. Therefore, it can be seen as a constrained rectangle packing problem, but for which the number of rectangles to be used and their size are not known a priori. This paper proposes an efficient way of solving this problem using constraint programming. The model is based on an open variant of the DiffN global constraint in order to deal with an unfixed number of rectangles, as well as a simple but efficient search procedure to solve this problem. An empirical evaluation shows the practical impact of every choice in the design of our model. A prototype implemented in the general purpose solver Choco is intended to assist architect decision-making in the context of building thermal retrofit.

Towards a BIM Approach for a High Performance Renovation of Apartment Buildings

Building renovation is a key issue for energy saving in the future. The goal of this article is to show how this traditional activity can be industrialized thanks to a BIM supported process. The critical needs are explained and the proposed BIM presented.