Ph. D. Project
Collaborative and multi-level planning of construction sites based on Lean and BIM
2021/11/26 - 2024/11/25
The ISoBIM ANR project (Proposal for a collaborative process for retrofitting by external Insulation
based on Lean and BIM paradigms) aims to provide an original response to the development of energy
renovation activities. At the same time, it seeks to help SMEs in the wood construction trade to achieve
digital transition and to improvetheir productivity. The ISOBIM proposal is innovative. On the one hand,
because it seeks to cover the entire renovation process from the identification of the constructive
solution, through the elaboration of configuration and layout models, to the elaboration of planning
models and monitoring of construction projects. And, on the other hand, because of its holistic and
anthropocentric approach, which is part of a break with scientific research in the field of control and
management of industrial/building systems. This dissertation will focus on the sub-process of planning
and scheduling of construction site activities. The large size of the panels (from a few m² to more than
30 m²) as well as their mass (from 50 kg to a few tons) pose problems of planning and scheduling at
finite capacity of critical resources such as cranes, storage areas or means of transport for the lifting of
these panels. Thus, the problem of planning and scheduling renovation activities on a building site can
be seen as a classic Resource Constrained Project Scheduling Problem (RCPSP) which consists in
minimizing the duration of a project by satisfying constraints of precedence and availability of
resources. However, in the context of renovation, activities are subject to multiple disruptions that can
degrade project performance: poorly estimated task durations, unavailability of unplanned resources,
weather conditions, etc [13], [14]. Two directions can be explored: a response to these uncertainties by
(i) the establishment of a robust plan or by (ii) a dynamic re-sequencing based on the real constraints of
the construction site. Considering the holistic and anthropocentric approach defended in this project,
the path that we wish to explore is the use of the Last Planner System. The implementation of the LPS
planning process is based on the animation of four schedules (two long term Master Scheduling and
Phase Scheduling) and (two short term Look-Ahead Planning and Weekly Work Planning). For the first
two schedules, planning is deliberative and results in the definition of important project milestones.
This is where the need for robustness is most important. For the last two, planning (in the sense of
scheduling) is collaborative and relies on the commitments of the last planner (the one who executes
the task) to eliminate or verify the satisfaction of the actual execution constraints. This is where the
dynamic re-sequencing aspect can play its full role. The review of the literature highlights some
limitations in the definition of the scheduling structuring parameters, but above all the risk of
desynchronization of schedules due to the non-propagation of constraints [35]. Objective: The
implementation of an LPS therefore implies the design of a model in several plans, each plan being the
place for a set of decisions of specific periodicities and horizons. This step is crucial because inadequate
temporal precision and frequency of rescheduling can lead to a desynchronization of the different
schedules. In addition, the structuring of the plans, the identification of constraints and their modeling
as well as the implementation of performance indicators for measuring the reliability of the schedules
are not trivial. Moreover, they are often only based on business practices and planners' knowhow
(which is one of the barriers to the adoption of LPS). A first objective will be, on the basis of an indepth
literature study, to propose a formalization of these good practices and a methodological guide for the
implementation of the LPS, in particular, for ISOBIM's target sites. Moreover, for several years, research
on hierarchical planning structures [15], [16] has shown the importance of the principle of
disaggregating decisions from one level to another. A poor coordination of planning objectives or the
propagation of constraints between levels will also lead to a desynchronization of schedules and
consequently a less
efficient planning. The study of the literature on LPS shows that no explicit link between the lower
levels of the plan and the higher level is formalized. To keep the principle of LPS autonomy while
guiding the actors of the construction site in their decision making, a first track will be to develop
interactive optimization models under constraints at each level and to couple them with BIM4D
simulation tools to validate the decisions of the level in question and their impacts on the schedules of
the higher levels. The development of these optimization models and the definition and specification of
the simulation models will be the second and main objective of this thesis.
BIM4D, LPS, Scheduling
Modeling and Control of Industrial Systems