Balancing large-scale machining lines with multi-spindle heads using decomposition

The paper deals with optimal logical layout design for a type of machining lines. Such lines are made of many machine-tools (workstations) located in sequence. On each workstation there are several multi-spindle heads. A spindle head does not execute one operation but a block of machining operations simultaneously. The problem studied in this paper consists of finding the best partition of the set of all operations to be executed on the line into blocks and workstations. The objective is to minimize the number of blocks and workstations. An optimal decision must satisfy a desired productivity rate (cycle time) and precedence and compatibility constraints for machining operations. A heuristic approach based on decomposition of a Mixed Integer Programming (MIP) model is developed. Two ways of forming sub-problems are proposed. One treats the obtained subsets independently. The second aggregates the solution of the previous subproblems. Results of their computational evaluation are reported.