Project scheduling: new approaches and solving methods

• Autores: Sofía Rodríguez Ballesteros
• Directores de la Tesis: Javier Alcaraz Soria (dir. tes.) , Laura Antón Sánchez (codir. tes.)
• Lectura: En la Universidad Miguel Hernández de Elche ( España ) en 2025
• Idioma: español
• Tribunal Calificador de la Tesis: Eva Vallada Regalado (presid.) , Marina Leal Palazón (secret.) , Francisco Saldanha Da Gama (voc.)
• Texto completo no disponible (Saber más ...)
• Resumen

  • This doctoral thesis delves into the project scheduling literature, a core field within operations research encompassing a variety of models and methods aimed at organizing interdependent activities over time---often subject to precedence and resource constraints---while optimizing specific objectives. As modern project environments become increasingly dynamic and uncertain, traditional models struggle to reflect key features such as multi-objective formulations, time-dependent costs, alternative execution modes, and incomplete information at the planning stage. This research addresses these limitations by proposing new mathematical formulations and tailored solution methods that improve the realism, flexibility, and effectiveness of scheduling models, ultimately contributing to more resilient and evidence-based decision-making.

    Three interrelated research lines are developed, each addressing a critical aspect of project scheduling under realistic constraints. The first focuses on an extension of the classical resource-constrained project scheduling problem that incorporates time-varying resource costs, aiming to jointly minimize project duration and total cost. Several metaheuristic algorithms are adapted and rigorously evaluated through computational experiments, revealing performance trade-offs and confirming the effectiveness of evolutionary strategies in producing diverse, high-quality solutions. The second line expands the model by integrating time-varying resource capacities and multiple execution modes per activity, each reflecting different trade-offs between resource use and duration. A tailored metaheuristic with problem-specific operators is proposed, demonstrating robustness and scalability, consistently outperforming exact methods on medium- and large-scale instances while remaining competitive on smaller ones. The third line introduces a scheduling problem with starting-time-dependent costs, defined within a budgeted uncertainty set. Motivated by applications with fixed execution order and flexible starting times, a robust two-stage optimization framework is proposed. A compact mixed-integer formulation is introduced for the continuous case, while an iterative algorithm is developed for the discrete counterpart. Computational experiments confirm compact the model's efficiency, scalability, and ability to deliver high-quality solutions under uncertainty.

    Overall, this thesis advances project scheduling by offering flexible modeling frameworks and efficient solution methods that bridge the gap between theory and practice. The proposed approaches support multi-objective decision-making, incorporate execution flexibility, and yield reliable schedules in uncertain environments. Together, these contributions provide valuable tools that enhance planning performance in complex, real-world scenarios.