The efficient planning of assembly lines is a central topic in production management. In practice, however, the skills and execution times of employees differ – not every person can perform every task equally quickly or at all. The Assembly Line Balancing with Worker Assignment Problem (ALBWAP) extends the classical Assembly Line Balancing Problem (SALBP) by this realistic dimension and thereby increases the complexity considerably.
Traditionally, the problem is modeled as a Mixed-Integer Program (MIP) and solved for example with Gurobi. With Hexaly, however, a new solver is available that employs its own approach, the so-called Set-based Modeling. Instead of a classical MIP formulation with binary variables and linear restrictions, the model in Hexaly is described using sets, relations, and constraints. Based on this, Hexaly combines different exact and heuristic search methods.
For the simple SALBP, Hexaly has already published benchmark results. These show that for large problem instances a significantly smaller gap to the optimal solution is achieved than with Gurobi. For the extended ALBWAP, however, a comparative approach with a detailed analysis of solution and instance structures is still missing. This gap shall be closed with this thesis.
As part of the Master’s thesis, the ALBWAP shall be modeled and solved both as a MIP (Gurobi) and as a Set-based Model (Hexaly). Afterwards, both paradigms shall be systematically compared based on different criteria. For this purpose, benchmark instances will be developed by extending existing SALBP instances with heterogeneous employees with individual skills. Optionally, a simple heuristic may additionally be implemented in order to classify the solver results for large instances.
Programming skills and initial experience with solvers (e.g., Gurobi, CPLEX, Hexaly) are advantageous.
If you are interested, please contact Judith Schulze.