Equilibrado de líneas de montaje en paralelo con estaciones multilínea y dimensionado de buffers
(English) Assembly lines are mass production systems which are relevant in the manufacture of standard and customized products. One of the most elementary optimization problems in this field is the Assembly Line Balancing Problem (ALBP). ALBP consists of assigning a set of tasks to a set of ordered...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis doctoral |
| Fecha de publicación: | 2022 |
| País: | España |
| Institución: | Universitat Politècnica de Catalunya (UPC) |
| Repositorio: | UPCommons. Portal del coneixement obert de la UPC |
| Idioma: | español |
| OAI Identifier: | oai:upcommons.upc.edu:2117/380365 |
| Acceso en línea: | https://hdl.handle.net/2117/380365 https://dx.doi.org/10.5821/dissertation-2117-380365 |
| Access Level: | acceso abierto |
| Palabra clave: | Àrees temàtiques de la UPC::Economia i organitzacio d'empreses |
| Sumario: | (English) Assembly lines are mass production systems which are relevant in the manufacture of standard and customized products. One of the most elementary optimization problems in this field is the Assembly Line Balancing Problem (ALBP). ALBP consists of assigning a set of tasks to a set of ordered stations, satisfying some specific constraints, in such a way that one or more objectives are optimized. Nowadays, companies dedicated to production usually have more than one assembly line or multiple lines to face demand fluctuations, design reasons, group products in different lines, etc. The use of multiple lines located in parallel has attracted the interest of researchers in the last years. Parallel assembly lines are two or more lines that, if they are close enough, can be balanced together using shared stations (multi-line stations) between adjacent lines. In a multi-line station, the operator performs tasks assigned to his/her station of two adjacent lines in the cycle time of each line. Parallel Assembly Lines Balancing Problem (PALBP) can have lines with different cycle times. As a consequence of working with different cycle times, the systems with multi-line stations may have to produce in batches. In that case, the use of buffers in multi-line stations may be needed. The review of the state-of-the-art of PALBP reveals that no studies consider the use of some type of buffers in the case of production in batches. It should be noted that not considering the need (existence and sizing) of buffers in the PALB with lines with different cycle times can lead to the design of academic solutions that cannot be implemented in the industry. This doctoral thesis deals with the parallel assembly lines balancing problem and buffer sizing (PALBP-B). Precisely, PALBPB consists of a PALB system with multi-line stations and lines with different cycle times, in which there is the possibility of needing buffers. The problem of study in this thesis focuses on the PALBP-B of straight lines and single model. The PALBP-B resolution consists of two (sub)problems that should be solved simultaneously: 1) line balancing: assigning tasks to stations (and its definition as regular or multi-line stations); and 2) buffers sizing in the multi-line stations. The review of the state-of-the-art reveals that the problem presented in this thesis has not been studied in the literature. This implies defining and formalising the PALBP-B. First, the buffer sizing (sub)problem is characterized and defined, and a mixed integer linear programming (MILP) model and two modifications derived from it are presented for its optimal resolution. In addition, non-exact procedures (heuristics and metaheuristic) are also presented to solve real-size instances. Secondly, nonexact procedures (heuristics and metaheuristics) are presented in order to solve the PALBP-B, that is, the line balancing (sub)problem and the buffer sizing (sub)problem (using the procedures developed previously for this purpose) together. All the methods developed to solve the buffer sizing (sub)problem are evaluated via a computational experiment based on a set of instances generated for this purpose. The approximate approaches developed to solve the PALBP-B are evaluated via an exhaustive computational experiment based on a set of realistic instances from the literature. |
|---|