TSALBP: Time and Space constrained Assembly Line Balancing Problems
A robustness information and visualization model for time and space assembly line balancing under uncertain demand
Authors
Manuel Chica2, Óscar Cordón2, Sergio Damas2 and Joaquín Bautista1
2 European Centre for Soft Computing
Abstract
The time and space assembly line balancing problem (TSALBP) is a realistic multiobjective version of assembly line balancing industrial problems involving the joint optimization of conflicting criteria such as the cycle time, the number of stations, and the area of these stations. However, the existing problem formulation does not consider the industrial scenario where the demand of a set of mixed products is variable and uncertain. In this work we propose to introduce novel robustness functions to measure how robust the line configuration is when the production plans demand changes. These functions are based on the stations overload under future demand conditions and are used as additional a posteriori information for the non-dominated solutions found by any multiobjective optimization method. The values of the robustness functions are put together with a novel graphical representation to form a generic model that aims to offer a better picture of the robustness of the set of Pareto-optimal solutions.
Real data from the assembly line and production planning of the Nissan plant of Barcelona is considered for the experimentation. This information is also employed to develop a new TSALBP instance generator (NTIGen) that can generate problem instances having industrial real-like features. The use of the robustness information model is illustrated in an experimentation formed by a set of instances generated by NTIGen. Results show how the use of this robustness information model is necessary for the decision maker as it allows her/him twen different assembly line configurations when future demand conditions vary.
Key words: Robust Optimization, Visualization, Time and Space Assembly Line Balancing, Mixed Products, Uncertain Demand, Problem Instance Generator
