| Job shop scheduling problems in general fall under a class of NP-hard problems. For a manufacturing service company operating under a job shop environment, the task of scheduling labor and jobs to the machines is extremely complicated. This can be attributed to the many factors involved in it including sequence dependent set-ups, maintenance schedules, variable processing times, uncertain demand, job due dates, flexible labor shift schedules, labor overtime, multiple parallel machines, labor dependent machine operability, and variable labor requirements.;In this study, a recursive routine was developed to determine a feasible job and labor schedule, aiming to minimize the cost of weighted tardiness, cost of operating machines, and cost of labor overtime. To search for a near optimal solution to the dual objective problem, two simulated annealing based algorithms were encoded in PHP. A MySQL based database was split in to training-data used to determine the algorithm parameters, and the variable demand instance test-data to run cost trials. For the purpose of achieving visibility, altering the MySQL database, and dispersing updated schedules, a dynamic web-based client-side user interface encoded using HTML5, CSS, JavaScript, JQuery, Ajax, and PHP was developed.;The results of the trial runs on the training data suggested that for the first algorithm, temperature reduction factor had a significant effect on the execution time, but not in the objective. The starting temperature, on the other hand, had no effect on either. On the contrary for the second algorithm, the temperature reduction factor and the starting temperature, had a significant effect on both the execution time and the tardiness objective. Further analysis suggested that increasing the total number of employees or multi-tasking of employees was essential to reduce the total tardiness. When the objective was altered to minimize the total cost as opposed to the total tardiness, it was observed that the tardiness weight was instrumental in reducing the total tardiness. This effect was true only upto a certain increased weight beyond which it did not affect the total tardiness. Similarly, an inverse relation was noticed between the total overtime and the total tardiness. Higher the overtime, lower the tardiness.;The concept of using problem specific information, and exploiting the dependency relation of the different elements to simplify the problem, proved an effective strategy for solving a multi-dimensional problem. |
