Cyclical scheduling and assembly line balancing

Abstract

This thesis addresses extensions of the assembly line balancing problem tied to cyclical scheduling and proposes mathematical models solution methods tackle such problems. Each of the thesis main chapter is based on a specific publication by the author throughout his doctorate. The main contributions include: mathematical formulations that can describe the steady state of mixed-model assembly lines when the product sequence is cyclical; a decomposition to solve the balancing problem when product sequencing and buffer allocations are also decision variables; a simheuristic to balance mixed-model lines when the product sequence is stochastic; corrections to a previous multi-manned balancing article (by another author); a paradigm shift for multi-manned lines that allow shorter line lengths, along with an efficient decomposition and lower bounds for the resulting problem; a description of internal storage costs tied to fractional task allocations in the context of dynamic line balancing or work-sharing; a method to solve multi-objective problems with part-wise linear Pareto fronts, and a case that allows comparing different line control systems in the context of cyclical product sequences. In several chapters, comparisons to literature benchmarks demonstrate the superiority of the proposed methods and formulations. In others, analytical descriptions and contributions answer research questions related to important managerial considerations such as the bowl phenomenon, internal storage costs, and the influence of line control types to line performance.