This PhD thesis contains the research carried out during my PhD program in collaboration with Ferrarelle Spa, a primary Italian mineral water bottling company. This type of PhD program is intended for employees of private companies, who continue to work full time in the company during the PhD program, with maintenance of salary. The work was done under the supervision of Prof. Luca Bertazzi, Professor of Operations Research at the University of Brescia, and with the contribution of Prof. Giovanni Pantuso, Associate Professor of Operations Research at the University of Copenhagen, in the development of the second problem studied in the thesis. The topic of this thesis is the use of mathematical models for the optimization of production and logistics problems. In Operations having scientific tools to support managerial decisions is increasingly important for making company activities more efficient. This thesis deals with two optimization problems related to Production and Logistics, and it is organized as follows. Chapter 1 provides a review of the state of art of the literature concerning the problems under analysis. In Chapter 2, we study one of the most important problems in warehouse management: The Stock Allocation Problem, i.e., the problem to determine how to allocate the initial stock and the quantity produced to bins, and then how to manage picking operations from these bins. The objective is to minimize the total number of bins used, to optimize the use of the available storage capacity. For this problem, we formulate an integer linear programming model, prove that it is NP-hard, and solve it to optimality. Then, we design a tailored heuristic algorithm, inspired by the current rule of thumb used by the company. An extensive computational experiment allows us to show that this problem can be solved to optimality in a reasonable computational time on real-world instances, and that the heuristic provides near–optimal solutions, that are able to significantly improve the solutions used by the company. In Chapter 3, we provide an innovative approach to one of the most important goals in Make-to-Stock production systems: to have an optimal production plan, able to achieve at best the stock target levels of several products, to manage the uncertainty of the demand, and to take into account operational production constraints, such as minimum production quantities and setup operations. We propose an integrated Goal Programming and Chance Constrained Optimization approach, where Goal programming is used to balance the different goals in terms of stock target levels. Demand uncertainty is handled by using Chance Constrained Optimization. We first formulate a model at the tactical level, in which a monthly time unit is used. This model is applied in a rolling horizon way, for each month over a planning horizon of one year. Then, we formulate a model at the operational level, in which a daily time unit is used, able to take into account the specific operational production constraints. To validate the effectiveness of the proposed approach, we apply it to the real case provided by the mineral water bottling company. The comparison of the optimal solutions with the company solutions allows us to demonstrate the effectiveness of the proposed approach, in terms of ability of balancing the achievement of the stock target levels, cost minimization, and saturation of the production capacity. Finally, we report the conclusions drawn from the study presented. In summary, this thesis aims to be a valid contribution to the challenging task for academic and industrial research to address large-scale, complex optimization problems under various uncertainties. The studies presented and their approach aim to build a “bridge” between scientific research and its application to real business optimization problems.
Questa tesi di PhD contiene la ricerca svolta in collaborazione con Ferrarelle Spa, primaria azienda italiana di imbottigliamento di acqua minerale. Questa tipologia di PhD è destinata ai dipendenti di aziende private, i quali continuano a lavorare a tempo pieno in azienda durante il PhD, con mantenimento della retribuzione. Il lavoro è stato svolto sotto la supervisione del Prof. Luca Bertazzi, Professore Ordinario di Ricerca Operativa presso l’Università di Brescia, e con il contributo del Prof. Giovanni Pantuso, Professore Associato di Ricerca Operativa presso l’Università di Copenhagen, nello sviluppo del secondo problema studiato nella tesi. L’argomento di questa tesi è l’utilizzo di modelli matematici per l’ottimizzazione di problemi di Produzione e Logistica. Nell’area Operations disporre di strumenti scientifici a supporto delle decisioni manageriali è sempre più importante per efficientare le attività aziendali. Questa tesi affronta due problemi di ottimizzazione legati alla Produzione ed alla Logistica, ed è organizzata come segue. Il Capitolo 1 è dedicato alla revisione dello stato dell’arte della letteratura riguardante i problemi oggetto di analisi. Nel Capitolo 2 studiamo uno dei problemi più importanti nella gestione del magazzino: l’allocazione delle scorte, ovvero determinare come allocare lo stock iniziale e la quantità prodotta nelle corsie di stoccaggio disponibili, e come gestire l’attività di prelievo da queste corsie. L’obiettivo è ridurre al minimo il numero totale di corsie utilizzate, ottimizzando l’utilizzo della capacità di stoccaggio disponibile. Per questo problema, formuliamo un modello di programmazione lineare intera, dimostriamo che è NP-hard e lo risolviamo all’ottimo. In seguito, proponiamo un algoritmo euristico, ispirato all’attuale regola empirica utilizzata dall’azienda. I risultati computazionali ci consentono di dimostrare che questo problema può essere risolto all’ottimo, con un tempo di calcolo ragionevole su istanze reali, e che l’euristica fornisce soluzioni vicine all’ottimo, in grado di migliorare significativamente le soluzioni utilizzate dall’azienda. Nel Capitolo 3, proponiamo un approccio innovativo a uno degli obbiettivi più importanti nei sistemi di produzione Make-to-Stock: avere un piano di produzione ottimale, in grado di raggiungere al meglio i livelli target di stock, di gestire l’incertezza della domanda e di tenere conto dei vincoli operativi. Proponiamo un approccio integrato di Goal Programming e di Chance Constrained Optimization, utilizzando Goal Programming per bilanciare i diversi obiettivi in termini di livelli target di stock e Chance Constrained Optimization per gestire l’incertezza della domanda. Formuliamo in primo luogo un modello a livello tattico utilizzando un’unità di tempo mensile. Questo modello viene applicato con un approccio Rolling Horizon, per ciascun mese su un orizzonte di pianificazione di un anno. In seguito, formuliamo un modello a livello operativo, utilizzando un’unità di tempo giornaliera e tenendo conto degli specifici vincoli operativi. Per validare l’efficacia dell’approccio proposto, lo applichiamo al caso aziendale. Il confronto delle soluzioni ottimali con le soluzioni aziendali consente di dimostrare l’efficacia dell’approccio proposto, in termini di capacità di bilanciare il raggiungimento dei livelli target di stock, minimizzazione dei costi e saturazione della capacità produttiva. Riportiamo infine le conclusioni tratte dallo studio presentato. Questa tesi vuole essere un valido contributo all’impegnativo compito della ricerca accademica ed industriale di affrontare problemi di ottimizzazione complessi e su larga scala in caso di incertezza. Gli studi presentati e il loro approccio mirano a costruire un “ponte” tra la ricerca scientifica e la sua applicazione in problemi reali di ottimizzazione aziendale.
Optimization of Production and Logistics: Mixed Integer Linear, Goal and Chance Constrained Programming Models.
Pedersoli, Felice
2025
Abstract
This PhD thesis contains the research carried out during my PhD program in collaboration with Ferrarelle Spa, a primary Italian mineral water bottling company. This type of PhD program is intended for employees of private companies, who continue to work full time in the company during the PhD program, with maintenance of salary. The work was done under the supervision of Prof. Luca Bertazzi, Professor of Operations Research at the University of Brescia, and with the contribution of Prof. Giovanni Pantuso, Associate Professor of Operations Research at the University of Copenhagen, in the development of the second problem studied in the thesis. The topic of this thesis is the use of mathematical models for the optimization of production and logistics problems. In Operations having scientific tools to support managerial decisions is increasingly important for making company activities more efficient. This thesis deals with two optimization problems related to Production and Logistics, and it is organized as follows. Chapter 1 provides a review of the state of art of the literature concerning the problems under analysis. In Chapter 2, we study one of the most important problems in warehouse management: The Stock Allocation Problem, i.e., the problem to determine how to allocate the initial stock and the quantity produced to bins, and then how to manage picking operations from these bins. The objective is to minimize the total number of bins used, to optimize the use of the available storage capacity. For this problem, we formulate an integer linear programming model, prove that it is NP-hard, and solve it to optimality. Then, we design a tailored heuristic algorithm, inspired by the current rule of thumb used by the company. An extensive computational experiment allows us to show that this problem can be solved to optimality in a reasonable computational time on real-world instances, and that the heuristic provides near–optimal solutions, that are able to significantly improve the solutions used by the company. In Chapter 3, we provide an innovative approach to one of the most important goals in Make-to-Stock production systems: to have an optimal production plan, able to achieve at best the stock target levels of several products, to manage the uncertainty of the demand, and to take into account operational production constraints, such as minimum production quantities and setup operations. We propose an integrated Goal Programming and Chance Constrained Optimization approach, where Goal programming is used to balance the different goals in terms of stock target levels. Demand uncertainty is handled by using Chance Constrained Optimization. We first formulate a model at the tactical level, in which a monthly time unit is used. This model is applied in a rolling horizon way, for each month over a planning horizon of one year. Then, we formulate a model at the operational level, in which a daily time unit is used, able to take into account the specific operational production constraints. To validate the effectiveness of the proposed approach, we apply it to the real case provided by the mineral water bottling company. The comparison of the optimal solutions with the company solutions allows us to demonstrate the effectiveness of the proposed approach, in terms of ability of balancing the achievement of the stock target levels, cost minimization, and saturation of the production capacity. Finally, we report the conclusions drawn from the study presented. In summary, this thesis aims to be a valid contribution to the challenging task for academic and industrial research to address large-scale, complex optimization problems under various uncertainties. The studies presented and their approach aim to build a “bridge” between scientific research and its application to real business optimization problems.File | Dimensione | Formato | |
---|---|---|---|
Tesi Pedersoli Corretto.pdf
accesso aperto
Dimensione
10.66 MB
Formato
Adobe PDF
|
10.66 MB | Adobe PDF | Visualizza/Apri |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/190463
URN:NBN:IT:UNIBS-190463