Embedding Ensemble Learning into Construction Optimisation : A Computational Reduction Approach

• Simulation-based optimisation (SO), which combines simulation and optimisation technologies, is a popular approach for construction planning optimisation. However, in the framework of SO, the simulation is continuously invoked during the optimisation trajectory, which increases the computing loads to levels that are unrealistic to support the real-time construction decision. This study proposes ensemble learning embedded simulation optimisation (ESO) as an alternative approach for construction optimisation. The ensemble learning (EL) algorithm modifies the SO framework through establishing a connection between the simulation and optimisation technologies. This approach reduces the computing loads associated with the conventional SO framework by accurately learning from simulations and providing efficient fitness evaluations for optimisation. A large-scale project application shows that the proposed approach was able to reduce the computing loads of SO by approximately 90% yet still provide comparable optimisation quality. The proposed method is an alternative approach to SO that can be run on standard computing platforms and supports nearly real-time optimisation decisions.

Subject headings

TEKNIK OCH TEKNOLOGIER  -- Samhällsbyggnadsteknik -- Byggproduktion (hsv//swe)

ENGINEERING AND TECHNOLOGY  -- Civil Engineering -- Construction Management (hsv//eng)

Keyword

Ensemble learning

Construction simulation

Decision optimisation

Construction method

Byggproduktion och teknik

Construction Management and Building Technology

Publication and Content Type

ref (subject category)

art (subject category)