Pavement rehabilitation, especially in a highway context, comprises a complex set of operations from milling the degraded pavement to paving the new layers, while also ensuring the traffic passage. The importance of having serviceable and safe highway networks makes this type of operations very time sensitive, since this usually involves taking over traffic lanes, inevitably causing delays to users, and in extreme situations end up in bottlenecks.
On top of that, this construction type also relies on very expensive heavy mechanical equipment. Even though there is an increasing interest in completing these rehabilitation interventions with the lowest costs and durations, the planning of this process is currently mostly based on the designer’s experience and little has been done to optimize it.
Therefore, the present study promoted by the BUILT CoLAB team, Margarida Amândio and Manuel Parente, together with Prof. José Neves, addresses the development of an intelligent optimization system based on an evolutionary multi-objective approach (NSGA-II) capable of supporting decision making in the planning of asphalt pavement rehabilitation interventions from the designer’s viewpoint.
By establishing a parallel to the manufacturing industry, rehabilitation processes such as these can be interpreted as a series of production lines. Hence, the developed system seeks to simultaneously minimize time and cost, inherent to the resources that compose those production lines, by finding the best resource allocation solution for the entirety of the rehabilitation process.
Implementation on a real highway rehabilitation case study demonstrated that the system is able to provide several optimal resource allocations solutions, and even discover new alternatives that were not originally considered.