Preprocessing rules for the dynamic layout problem
The dynamic layout problem model takes into consideration future changes in a facility by minimizing material flow costs plus the cost of rearrangement. The dynamic layout problem is solved by first using a quadratic assignment problem and then using a mixed integer linear problem. The number of integer variables required by the mixed integer linear problem restricts the size of problem that can be solved.Preprocessing rules reduce the number of integer variables required by the dynamic layout problem using two methods. The first method eliminates redundant integer variables used to represent stationary department pairs. The second method uses the layout generated by the quadratic assignment problem, the material flow matrices and the department sizes to predict the values of the integer variables. The preprocessing rules are implemented within the two-stage algorithm, before execution of the mixed integer linear program. Preprocessing rules allow problems with up to twenty departments and five time periods to be solved. Preprocessing rules decrease the total cost of a layout by an average of 16.95% compared to published solutions.
School Location:USA - Ohio
Source Type:Master's Thesis
Keywords:preprocessing rules dynamic layout problem
Date of Publication:01/01/1994