Transforming Construction Through Smart Logistics

Faced with the housing crisis, labor shortages, and skyrocketing costs, the construction industry no longer has the luxury of doing "business as usual." For Amin Chaabane, a professor and researcher at ÉTS specializing in logistics and supply chains, change is no longer optional: construction must become more efficient by drawing inspiration from the manufacturing sector, while respecting the very real constraints of construction sites.

“We are now seeing the emergence of logistical challenges in construction that we were already acutely familiar with in industrial manufacturing,” he explains. Detailed planning, task scheduling, and flow synchronization are well-known concepts in industrial engineering, but they remain difficult to implement in a sector riddled with unexpected events.

When construction becomes (somewhat) industrial

Construction is increasingly moving towards some form of industrialization, especially off-site construction. Panels, modules, or components are built in a factory, then shipped and assembled on site. The payoff is higher quality, shorter delays, and tighter cost control. These are crucial advantages in  dense urban areas and in northern or Arctic regions, where harsh weather makes on-site work complicated.

But not everything can be industrialized. “Some processes lend very well to off-site production, others do not,” Amin Chaabane points out. This hybridization makes planning far more complex: factory production, transportation, handling, municipal constraints (module size, site access), subcontractor availability, and on-site conditions must all be coordinated at once.

Planning: A real-time puzzle

On a construction site, unforeseen events are the norm: supplier delays, equipment breakdowns, volatile weather conditions, and last-minute changes. Without the right tools, planning becomes nearly impossible when several projects are running in parallel.

This is where Amin Chaabane's research comes in. His work draws on advanced optimization and planning methods, fueled by rich data collected in the field. For more than a decade, he has been developing models that operate across multiple levels (strategic, tactical, and operational) and integrate tools such as building information modeling (BIM) to improve coordination between stakeholders.

Artificial intelligence: Predicting for better adaptation

More recently, the professor has been exploring how artificial intelligence can make planning more proactive. AI enables the creation of predictive models that analyze construction site data to detect potential anomalies, such as delivery delays, congestion risks, and coordination issues.

These predictions are then sent to an optimization engine that automatically adjusts the schedule. However, the goal is not to constantly optimize everything. “We must avoid over-optimization,” emphasizes Professor Chaabane. The challenge is to find the right balance between stability and performance, so that advanced planning systems reamain usable in the field.

This approach, which combines optimization and machine learning, particularly deep reinforcement learning, enables a shift from rigid planning to dynamic management, allowing real-time responsiveness.

A more sustainable and better coordinated supply chain

Logistics is not limited to economic performance. Amin Chaabane wants to fully integrate sustainability into supply chains, taking into account economic, environmental, and social criteria, including social acceptability.

His work on the circular economy in the construction sector aims to reduce waste generated during renovation and demolition activities promoting the recovery, sorting, reuse, and recycling of materials instead of landfilling them. He has developed optimization models to improve the planning of waste collection, transportation, and treatment operations, while accounting for uncertainties related to waste volumes, material quality, and supplier or recovery facility locations. This research has been applied in Quebec and has contributed to the design and improvement of recovery and recycling networks, supporting more sustainable and efficient resource management practices.

Tools designed for people in the field

Another key aspect of Professor Chaabane’s research focuses on making advanced optimization tools accessible to non-experts. His goal is to develop AI-powered decision-support systems that eliminate the need for specialized knowledge in programming, mathematical modeling, or logistics.

By combining Generative AI, Agentic AI, and optimization techniques, engineers and project managers can interact with intelligent agents using natural language to describe objectives and constraints. These agents automatically formulate and solve optimization models, generating recommendations for procurement, subcontractor allocation, logistics, scheduling, and resource planning.

Beyond generating plans, the agents can monitor operations, assess alternative scenarios, and dynamically adapt decisions as conditions change. This approach enables predictive-reactive decision-making and supports a transition from static planning to intelligent, real-time management.

Tangible benefits for industry and society

Tested in collaboration with industry, this has delivered tangible benefits, including faster project execution, lower costs, improved site safety, increased operational predictability, and reduced environmental footprint.

More broadly, Professor Chaabane’s work is contributing to the development of a new generation of smart and sustainable supply chains for the construction sector. By integrating advanced optimization, artificial intelligence, and digital technologies, his research supports faster, more resilient, and environmentally responsible construction systems, helping address Canada’s growing housing and urban development challenges.