In the context of a rapidly growing e-commerce industry, particularly accelerated by the COVID-19 pandemic, and the changes in the customer behaviour produced since then, i.e. customers demanding low volume orders with high variety and expecting high service quality plus short delivery times, companies mut adapt their logistics to remain competitive in this new business arena. Moreover, in the recent years Industry 4.0 has been taking place. This new Industrial Revolution is based on, in general terms, connectivity, data analytics and data intelligence, computational power, advanced engineering and human-machine interaction. This study will explore a human-robot collaborative picking system with a swarm configuration through a simulation developed in Python. It consists in a picking system where robots consolidate orders, travels to the picking locations autonomously and wait for an available picker performs the picking activity, and, once finished the picking mission, the robot goes to the depot and then it continues operating until the end of the shift. The aim of this thesis is to analyse how the performance of this picking system (the average order throughput time) and the utilization of the agents (average robot utilization and average picker utilization) change when: the number of pickers and number of robots change while keeping a constant order arrival rate (RQ1 and RQ2), and the number of pickers, number of robots and the order arrival rate change while keeping the ratio between the number of pickers and the number of robots constant (RQ3 and RQ4). After running the simulation, the output is deeply explained using several plots and tables. It was observed that by increasing the number of pickers and robots the performance of the system improves, but the improvement produced by adding more pickers is more important, until certain point in which the number of pickers is enough to manage the order arrival rate. Also, increasing the order arrival rate produces a worse performance if keeping a fixed number of agents. Regarding the utilization of the agents, the increase in the order arrival rate increases both pickers and robots’ utilization. By augmenting the number of pickers, the pickers’ utilization decreases but it does not affect the utilization of the robots. Finally, increasing the number of robots slightly increases the utilization of the pickers and decreases the utilization of the robots.
L’industria dell'e-commerce è in rapida crescita, ed è stata particolarmente accelerata dalla pandemia, e dei cambiamenti nel comportamento dei consumatori che sono avvenuti da allora, ovvero clienti che richiedono ordini di volume ridotto ma con alta varietà e si aspettano alta qualità del servizio e tempi di consegna brevi. In questo contesto le aziende devono adattare la loro logistica per rimanere competitive. Inoltre, negli ultimi anni sta avendo luogo l'Industria 4.0. Questa nuova rivoluzione industriale si basa, in termini generali, sulla connettività, l’analisi dei dati e data intelligence, potenza computazionale, ingegneria avanzata e interazione uomo-macchina. Questo studio esplorerà un sistema di picking collaborativo uomo-robot con una configurazione swarm, attraverso una simulazione sviluppata in Python. Si tratta di un sistema di picking in cui i robot consolidano gli ordini, si spostano autonomamente verso le picking location e attendono che un picker disponibile esegua l’attività di picking, e una volta completata la missione di picking, il robot torna al deposito e continua a operare fino alla fine del turno. L’obiettivo di questa tesi è analizzare come le performance di questo sistema di picking (tempo medio di transito dell’ordine) e l’utilizzo degli agenti (utilizzo medio dei robot e utilizzo medio dei picker) cambiano quando: il numero di picker e robot cambia mantenendo costante il tasso di arrivo degli ordini (RQ1 e RQ2), e il numero di picker, robot e il tasso di arrivo degli ordini cambia mantenendo costante il rapporto tra il numero di picker e il numero di robot (RQ3 e RQ4). Dopo aver eseguito la simulazione, i risultati vengono spiegati in dettaglio attraverso vari grafici e tabelle. È stato osservato che aumentando il numero di picker e robot, le performance del sistema migliorano, ma il miglioramento prodotto dall’aggiunta di più picker è più significativo, fino a un certo punto in cui il numero di picker è sufficiente per gestire il tasso di arrivo degli ordini. Inoltre, l'aumento del tasso di arrivo degli ordini peggiora le performance se si mantiene un numero fisso di agenti. Per quanto riguarda l’utilizzo degli agenti, l'aumento del tasso di arrivo degli ordini aumenta sia l’utilizzo dei picker che quello dei robot. Aumentando il numero di picker, l’utilizzo dei picker diminuisce, ma non influisce sull’utilizzo dei robot. Infine, l'aumento del numero di robot aumenta leggermente l’utilizzo dei picker e riduce l’utilizzo dei robot.
Performance analysis of a human-robot collaborative-picking system with swarm-based configuration
SALINAS MAURER, DIEGO NICOLAS
2023/2024
Abstract
In the context of a rapidly growing e-commerce industry, particularly accelerated by the COVID-19 pandemic, and the changes in the customer behaviour produced since then, i.e. customers demanding low volume orders with high variety and expecting high service quality plus short delivery times, companies mut adapt their logistics to remain competitive in this new business arena. Moreover, in the recent years Industry 4.0 has been taking place. This new Industrial Revolution is based on, in general terms, connectivity, data analytics and data intelligence, computational power, advanced engineering and human-machine interaction. This study will explore a human-robot collaborative picking system with a swarm configuration through a simulation developed in Python. It consists in a picking system where robots consolidate orders, travels to the picking locations autonomously and wait for an available picker performs the picking activity, and, once finished the picking mission, the robot goes to the depot and then it continues operating until the end of the shift. The aim of this thesis is to analyse how the performance of this picking system (the average order throughput time) and the utilization of the agents (average robot utilization and average picker utilization) change when: the number of pickers and number of robots change while keeping a constant order arrival rate (RQ1 and RQ2), and the number of pickers, number of robots and the order arrival rate change while keeping the ratio between the number of pickers and the number of robots constant (RQ3 and RQ4). After running the simulation, the output is deeply explained using several plots and tables. It was observed that by increasing the number of pickers and robots the performance of the system improves, but the improvement produced by adding more pickers is more important, until certain point in which the number of pickers is enough to manage the order arrival rate. Also, increasing the order arrival rate produces a worse performance if keeping a fixed number of agents. Regarding the utilization of the agents, the increase in the order arrival rate increases both pickers and robots’ utilization. By augmenting the number of pickers, the pickers’ utilization decreases but it does not affect the utilization of the robots. Finally, increasing the number of robots slightly increases the utilization of the pickers and decreases the utilization of the robots.File | Dimensione | Formato | |
---|---|---|---|
Thesis_Salinas_v3.pdf
accessibile in internet solo dagli utenti autorizzati
Dimensione
2.03 MB
Formato
Adobe PDF
|
2.03 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/229980