Université Grenoble ALpes
Research Context
Planning the production of goods and services is an increasingly difficult task to optimize. Today, customers have many requirements in terms of quality, flexibility, availability and delivery time. Product diversity is increasing and demand can fluctuate sharply. Production processes are constantly evolving. Investments can be heavy and require justification, each material resource must be exploited to the fullest until its maximum potential is reached. For all these reasons, the use of production planning software is necessary. This software can be seen as a decision support tool. It allows a company, in theory, to achieve the objectives it has set for itself in the simplest possible way.
A production planning software should in theory be able to deal with the following issues:
- Forecasting needs in order to size production
- Optimization of investments: the choice and sequencing of investments to be undertaken are strategic decisions that make it possible to better size production units or equipment fleets;
- Optimization of equipment maintenance and life cycle: resource management by integrating the constraints of maintenance visits, equipment renewal at the right time (ageing of the fleet, fleet optimization), planning of maintenance tasks;
- Optimization of production resources: distribution of production on different means of production (including machines, operators), erasure, smoothing, detection and resolution of conflicts, etc.;
- Optimization of the production plan: assignment of production tasks to material resources, scheduling, optimal use of resources, optimization of energy contracts;
- Optimization of manufacturing processes: optimization of formulation or diversity.
- In practice, there are no tools capable of addressing all its optimization and planning issues. Production planning software is primarily a visualization tool. Optimization and planning tasks are still largely the responsibility of human experts.
PhD. Objectives
In this context, the recruited doctoral student will have to (1) develop a state of the art on management tools and techniques for industrial production planning; (2) develop an original prototype production planning software capable of integrating the various problems mentioned above by relying on techniques derived from artificial intelligence and in particular from automated planning. In artificial intelligence, automated planning or more simply planning, aims to develop algorithms to produce plans typically for execution by a robot or any other autonomous system. Planning software that incorporates these algorithms is called planners. The difficulty of the planning problem depends on the simplification assumptions that are taken for granted, such as atomic time, deterministic time, full observability, etc. A typical planner manipulates three inputs described in a formal language (such as PDDL) that uses logical predicates:
- a description of the initial state of a world,
- a description of a goal to be achieved and
- a set of possible actions (sometimes called operators).
Each action is specified by preconditions that must be met in the current state for it to be applied, and post-conditions (effects on the current state). The interest of planning in this context is to use a common descriptive language that allows production processes to be easily modelled, leaving the complexity of their optimization to a planner. To go further, have a look to http://pddl4j.imag.fr/.
Requirements:
The candidate must have:
- a Master 2 in Computer Science with successful research experience
advanced programming skills (design and implementation), especially in Java and C++ - a good academic level attesting to his/her ability to combine practice and theory
- a level of professional oral and written English
- general knowledge in the field of artificial intelligence
- an appetite for industrial issues
Application’s deadline: October 15, 2023.
General Information:
– Contract start date: from septembre 2023
– Duration of the contract : 36 months
– Working hours: 100% of the time
– Desired level of studies: Master 2
– Host laboratory: Laboratoire d’Informatique de Grenoble (UMR 5217) – Marvin Team
Procedure and contact
Send to Damien.Pellier@imag.fr:
- Your Master 2 diploma with your marks or just you marks if your are in Master 2
- Curriculum vitae with 2 references
- At least one letter of recommendation
- Your Master 2 internship thesis and any publications you may have
Applications are managed on a case-by-case basis. You will be informed promptly by email of the admissibility of your application and if you are invited to a first interview.