Abstract:

The field of scheduling metal surface treatment plants is extensively studied in the literature and it is referred to by the name of Hoist Scheduling Problem HSP. HSP is classified as flow-shop scheduling problem but differs because of the transportation constraints and the no-wait constraint. This research deals with a real anodizing plant in an extrusion company. In this plant a number of different profiles (jobs) are dipped in a series number of tanks which are filled with acids, water, dyes and other chemicals for varying lengths of time; the transportation of these jobs from one tank to another is done by five hoists (cranes). The plant is suffering from a main bottleneck in the hot sealing stage. The objective of this study to minimize the makespan of a demand of selected days and to overcome the main bottleneck problem. A heuristic algorithm was developed for scheduling the anodizing demand of the extrusion company. The algorithm was coded in Matlab (R2015b - editor) to simulate the production of certain days. The runs of the simulation were divided into three phases. In the first phase, the results were validated by comparing the anodizing times obtained by the algorithm with the actual times, the results showed a reduction in the makespan with an approximated average of 18%. In the second phase, different combinations were examined, where additional tanks were added and different job sequences were introduced to study the effect of these combinations on the makespan and on the main bottleneck problem. The results suggested that adding one additional hot sealing tank and sequence the jobs based on “Highest Recipe ID First” is the best combination. Additional jobs were added in the third phase to the actual number of jobs in each day using the actual configuration and two different sequences: one with the actual sequence and the other with the sequence that produced the minimum makespan. This is done in an attempt to utilize the remaining time that is left out after minimizing the makespans, the results revealed that the algorithm was able to utilize the time by an approximated average of 98% and to increase the number of jobs by an approximated average of 20%. What makes this study unique is it’s dealing with an anodizing plant thatconsiders loading, unloading stages and input and output buffers, something rarely seen in the literature where most of the studies are dealing with electro-plating plants and do not consider the loading and unloading stages.