Urban project planning for the large-scale urban system is a very hard problem.•
The holonic approach offers a proper view of the complex urban systems.•
In practice, greedy methods often move away from realistic solutions.•
Heuristic and metaheuristic approaches are well suited to tackle the BKP problem.
Hitherto, urbanization reached unprecedented spreading, various problems in the field increase from day to day, and makes the urban phenomena more dynamic and more complex. Therefore, it is important to call in experts and provide all stuff to establish urban projects’ plans, which often need to be achieved in a brief time. Actually, decision-makers need more and more updated plans and even sustainable solutions to convey eventual urban changes with maintaining intrinsic features of urban areas, such as coverage, inter-dependency, and coherency. Due to decision-makers yearnings and the short time allocated to planners, urban project planning remains an exhausting task; it resorts to arbitrary choices to find a good match of projects according to the intended situations. On the other hand, it should take care of the available resources like funds, land, water, energy, underground, and raw materials, which ought to be rationally exploited, and preserved for future generations. In this paper, the proposed intelligent decision support system (IDSS) aims to find out the best urban plans that fit urban projects to appropriate areas. It also employs the holonic approach to model complex and large-scale urban systems, where agents of each level apply a new multi-objective ant colony optimization algorithm called BKPACS for the urban project planning problem, which is viewed as a bounded knapsack problem (BKP). To produce global optimal urban plans, the main algorithm called H-MACO coordinates between the different levels of this holonic system. The experimental results on a set of urban projects about a province of Algeria show good quality plans produced in less time.