Source de subvention : XEROX PARC (USA), CRSNG (EGP), Trisotech Inc.
Professeurs impliqués: Hafedh Mili (LATECE), Renata de Carvalho (LATECE), Javier Gonzales-Huerta (LATECE)
Étudiants & stagiaires: Anis Boubaker
Résumé : It is 3:00 pm. You pop-up your 65th shipment order of the current shift. You look for books in the shelves, pick a box of the right size, put the books in the box. Print a shipping slip and an address label, put the slip in the box, close the box, paste the shipping label, and put the box on the conveyor belt. It is labor day week-end. Jogger A comes in complaining of chest pain. You measure their blood oxygen level. It is normal. No history of heart problems. You order an EKG to be done when the EKG machine is free. Ambulances brings too baseball players B and C who knocked heads at third base. Both have a royal headache, and B keeps fainting. You order an EKG for B, to be done ASAP, and CAT-scans for both. As you start connecting the just-freed EKG machine to B, ambulances bring in D and E who crossed paths on a biking trail. There is some bruising, bleeding, and perhaps fractured bones. While you try to figure out what to do with D and E, patient A faints in the waiting room. Skeptics argue that BPM and BPM systems (BPMS) have only solved the ‘easy problems’, such as our Amazon warehouse work: 1) business processes where a handful of contingencies handle all situations, and 2) processes where most of the activities can be automated. However, complex and knowledge intensive processes do not lend themselves to such rigid formalizations. In the ER example, the way a patient is handled depends on, 1) what brings them to ER (type of ailment and severity), 2) their medical history, and 3) what else is happening in ER at that time. ER staff have more or less well-defined protocols to follow, but are otherwise left to contend with, 1) tasks that are mostly manual (e.g. taking a blood sample) and knowledge intensive (diagnosis), 2) different instances of the same process that compete for scarce resources, and 3) the urgency to act. This project aims at developing languages and tools for the specification and execution of such processes so that, 1) the protocols and constraints inherent in such processes can be specified and enforced, without imposing a predetermined task flow, and 2) information and resource sharing can occur across competing instances of the same process.
Source de subvnetion: Fonds Québécois de Recherches en Nature et Technologies – projets en équipe
Professeurs impliqués : Ghizlane ElBoussaidi (LATECE), Yann-Gael Guéhéneuc (LATECE), Hafedh Mili (LATECE, PI), Naouel Moha (LATECE), Petko Valtchev (LATECE), Jean Privat (LATECE)
Étudiants & stagiaires: Manel Abdellatif (PhD, Polytechnique), Leith Gardabbou (DESS, UQAM), Anas Shatnawi (Postdoc, UQAM), Ellen Haas (MS, UQAM), Marie-Pier Lessard (BS, UQAM)
Résumé: L’architecture orientée services perçoit les applications d’affaires comme des orchestrations de services réutilisables déployés à travers l’organisation. Ce paradigme est né de la convergence d’un besoin d’affaires, et d’une évolution technologique permettant d’y répondre. Tant l’optimisation des processus d’affaires intra- et inter-organisationnels, que la mise en marché de nouveaux produits et services, se heurtent à l’hétérogénéité des applications existantes, et à leur faible potentiel de réutilisation. Des percées au niveau des normes et des architectures distribuées ont rendu possible le développement d’applications complexes par intégration de composants réutilisables, hétérogènes, et relativement indépendants, c-à-d, des services. Encore faut-il restructurer les applications existantes pour s’y conformer. Ce projet vise la restructuration architecturale orientée services d’applications légataires
orientées objet en, 1) analysant leur code pour y identifier des services potentiels, 2) encapsulant les fonctions ainsi identifiées dans des interfaces de services, et 3) restructurant le code existant pour rediriger les invocations de ces fonctions à travers ces interfaces. Notre projet est novateur parce que, 1) il vise à offrir une solution complète depuis l’identification des services dans une application légataire, jusqu’à la restructuration complète de cette application, 2) les algorithmes proposés tiennent compte de la complexité des applications légataires, d’une typologie fine de services, et des variantes technologiques de services, et 3) il met à profit les compétences complémentaires des membres de l’équipe en ré(tro)ingénierie de logiciels, en conception architecturale, en transformation de modèles, et en classification conceptuelle.
Source de subvention: CRSNG – Subvention à la découverte
Professeurs impliqués: Hafedh Mili, Javier Gonzalez-Huerta, Abdel Leshob
Étudiants & stagiaires: Anis Boubaker
Résumé: Organizations build information systems to support their business processes. One would expect organizations that use the same business processes to be able to use or reuse the same IT infrastructure. This may be true at the enterprise architecture level but does not translate into reuse at the more concrete software artifact level, where most of the IT resources are spent. There are many reasons for this. First, there is a wide variety of business processes for doing anything, from procurement, to logistics, to financials, which may share a common core, but differ in the detail. Second, for any given business process, there are different levels of IT support, ranging from a simple recording of the activities of an essentially human process, to full process automation. Model-driven development [BROW04], and software reuse in general, have achieved much progress in deriving software models at development stage n, from software models at stage n-1. They have done this either by codifying good solutions to recurrent problems (patterns, analysis, architectural, or design), or by codifying the transformations (from platform-independent/PIM to platform-specific models/PSM, or from PSM to code as supported by CASE tools), or both. However, much remains to be done for the transformation from business process models (as computation independent models, or CIMs) to analysis software models (PIM)s. This is the transformation that we propose to investigate in this work.
Source de subvention: CRSNG (SEP), Orckestra Inc (www.orckestra.com)
Professeurs impliqués: Hafedh Mili, Abdel Obaid
Étudiants & Stagiaires: Imen Benzarti, Anis Boubaker
Résumé : Imagine that you enter your favourite grocery store and are greeted with a flashing screen display: ”Bonjour [Cathy/Jean]“. In the meat section, you receive an SMS regarding the special on lamb chops that you are known to like. As you pick the lamb chops, you get another SMS recommending a Pinot Noir. Walking through the junk food isle, you drop a bottle of soda in the shopping cart, and get an SMS “Dr. Smith won’t like it!”. You walk by the diapers section and you are notified of a special on size 6 diapers because you have been buying size 5 diapers for the past three months.This is an example of a new generation of context-aware customer experience applications (CA-CEM) that take advantage of ubiquitous computing (the internet of things) and reams of data about customers and their shopping habits, to adapt a company (or organization)’s products and services to its customers’ needs/desires. Ethical issues notwithstanding, how to design such applications, from business decisions to technological implementations? What stages of the scenarios could benefit from IoT? What do these applications need about the customers? What functionalities should IoT vendors offer to retailers to implement their CA-CEM scenarios? The purpose of this project is to develop a methodology and a software framework to support 1) the business designof CA-CEM scenarios, 2) the software design of CA-CEM functionalities, and 3) the (partial) implementation of the designed CA-CEM functionalities.