This work aims to analyze and anticipate flood risks in areas downstream of the Bin El Ouidane and Ahmed El Hansali Dams, in a context marked by an increase in extreme weather events. The main challenge is to understand and model flood behavior in order to better assess vulnerable areas and reduce potential impacts on populations and infrastructure. 

To this end, an approach combining hydraulic simulation, historical data analysis, and spatial mapping was adopted. Simulations were carried out using HEC-RAS based on 44 years of rainfall data, followed by geospatial processing using QGIS to produce risk density maps. Two scenarios were studied: flooding and dam failure. 

The results obtained made it possible to identify the most exposed areas and establish a precise risk map. This work highlights the importance of simulation tools in preventive management and proposes perspectives for improving protection strategies and land-use planning.

Authors: Salma AADLAFI; Hafsa AHMOUCHI and Salah ATMANE

Supervisor: Mohamed EL ALAOUI

Faced with the recurring natural disasters that strike Morocco—the Al Haouz earthquake (magnitude 6.8; 2,960 victims; 2.8 million affected, September 2023), the floods in Safi and Ksar El Kébir (December 2025 – February 2026), and the winter isolation of more than 200 villages in the provinces of Béni Mellal, Azilal, and Khénifra—the Béni Mellal-Khénifra region suffers from a critical structural deficit in local logistics infrastructure. Response times are 48-96 hours due to the lack of decentralized warehouses, exposing 80,000 residents to life-threatening shortages. This work proposes a rigorous mathematical approach to determine the optimal number, location, and size of emergency storage centers in this region, based on real data from the RGPH 2024 (2,484 rural villages, 1,235,364 inhabitants spread over 17,555 km²). 

The problem is formulated as a Multi-Objective Capacitated Facility Location Problem (MO-CFLP), solved by the Gurobi solver via its Python API. The model simultaneously optimizes population-weighted distance, investment cost, and volumetric capacity, subject to constraints excluding risk zones (seismic, flood-prone, ecological, military), a maximum coverage of 100 km, a unique allocation of villages, and minimum capacity. Of 106 candidate sites selected after geospatial filtering (area ≥ 25 ha), four optimal centers were identified in the municipalities of Aït Abbas (Azilal), Aït Ishaq and Sbat Aït Rahou (Khénifra), and Ouled Youssef (Béni Mellal), ensuring comprehensive coverage of the 2,484 villages. A complete operational management model is also presented: internal architecture in four functional zones, Push/Cross-Docking flow strategies, last-mile optimization by VRP/TSP, and reverse logistics. 

Authors: Noha JAKIR; Imad EL ADES and Haytame EL ATRAOUI

Supervisor: Mohamed EL ALAOUI

The Travelling Salesman Problem (TSP) is one of the most studied combinatorial optimization problems in operations research. Despite its simple formulation, it belongs to the class of NP-hard problems, making its solution complex as the number of points to visit increases. In this work, we apply the TSP to a real-world industrial problem: optimizing the path of the drilling head during the fabrication of a printed circuit board (PCB). The objective is to determine the optimal order for visiting the holes in order to minimize the total distance traveled by the machine. This optimization reduces manufacturing time, energy consumption, and production costs. To solve this problem, two metaheuristics inspired by natural phenomena were studied: Grey Wolf Optimizer (GWO) and Symbiotic Organism Search (SOS). These algorithms were adapted to the discrete nature of the TSP using Random Key Encoding (RKE). The experiments performed show that both methods yield significant improvements. The best results were obtained with the SOS algorithm, which achieved a minimum distance of 1341.30 mm compared to 1397.45 mm for GWO under the same experimental conditions. An economic analysis also estimated a potential annual gain exceeding $73,000 for large-scale industrial production. 

Authors: Souhail KBIRI; Basma MOUINI and Mohammed HIMMI 

Supervisor: Mostapha OULCAID