A MODIFIED VARIABLE NEIGHBORHOOD SEARCH ALGORITHM FOR ALTERNATIVE SUBGRAPHS ASSEMBLY LINE BALANCING PROBLEM EQUIPPED HUMAN - ROBOT COLLABORATION RESOURCES

• المؤلف الرئيسي: Puspadewi Sriamine, Kanshadia
• التنسيق: Final Project
• اللغة: Indonesia
• الوصول للمادة أونلاين: https://digilib.itb.ac.id/gdl/view/68681
• المؤسسة: Institut Teknologi Bandung

The increasing complexity of the production process urges manufacturing companies to increase their flexibility in the production line. However, the highly flexible human operator is not enough to accommodate the increasing trend of mass customization. Cobot is one of the leading technologies to support human-robot collaboration (HRC) in a production line, due to its flexibility in working alongside a human in a workstation. Human-robot collaboration-based production line allows a certain process to be done using different resources, which increases the complexity of line balancing problem called alternative subgraph assembly line balancing problem (ASALBP). The ASALBP-HRC-based solution also requires efficient computation time to ensure optimal production planning. Previous research has developed a variable neighborhood search (VNS) algorithm to generate a solution for the ALBP-HRC-based production line. However, the algorithm could not accommodate more complex problems such as ASALBP which requires selecting a route from several possible alternatives. The proposed VNS algorithm minimizes cycle time as the objective function for the ASALBP-HRCbased production line. Development of the algorithm uses the primary base of VNS structure including shaking, local search, and move with several added processes including operator switching and selecting alternative subgraphs. This proposed VNS algorithm, in form of Python programming language, was able to generate suboptimal solutions with efficient computation time. The algorithm generated -94.3% computation gap and 86.2% result gap between the analytical method and VNS tested with 9 datasets. In addition, the proposed research includes an analysis of significant parameters used to acquire the final solution including number of neighborhoods for all datasets and number of shaking process iterations for data with small number of tasks.