(145b) Coupling Hoist Scheduling and Job Queue Optimization | AIChE

(145b) Coupling Hoist Scheduling and Job Queue Optimization


Fu, J. - Presenter, Lamar University
Xu, Q. - Presenter, Lamar University
Zhao, C. - Presenter, Lamar University

There are thousands of electroplating and polymeric coating shops in the U.S. annually producing numerous workpieces for many pillar industries.  Hoist scheduling is commonly recognized one of the most important factors to improve their productivity.  Nowadays, one challenging problem is that a plating shop should have the capability to simultaneously manufacture multiple types of jobs, and satisfy different custom requests.  Common methods for this consideration include: i) a production line would be programmed to manufacture different types of job sequentially, i.e., processing the second type of job after fully completing the first type of job; ii) furnish multiple production lines and let one production line take care of a specific type of job.  The first way is easy to set up but will cause significant delay for some job processing tasks.  The second way can meet rapid manufacturing request but inevitably induce more investment and cost.  To trade off between productivity and investment cost, a third way is proposed to mingle different types of job production in one single production line.  Under the situation, the hoist is scheduled in a cyclic way to simultaneously handle multiple types of job so as to maximize the overall productivity.  Conceivably, not only the hoist scheduling should be optimized, the best job loading queue to the production line should also be optimized. 

In this paper, a systematic methodology coupling optimization of cyclic hoist scheduling and job loading queue for maximum productivity has been developed.  It is programmed as an MILP (Mixed Integer Linear Programming) model addressing all the major concerned hoist scheduling issues, such as cyclic scheduling for large amount of jobs with multiple recipes, job queue determining, various production unit capacities, strict processing time requirements and exact manufacturing proportion for further assembling.  This MILP model is programmed and solved in GAMS version 23.3 with the solver CPLEX.  Since the model is fully linearized, the global optimal solution can also be guaranteed.   The efficacy of the developed methodology is demonstrated by a case study for an electroplating system.  The comparison for the cyclic hoist scheduling with and without job queue optimization is also conducted.