Scheduling plays a crucial role in manufacturing industries. If a product has not been completed on time, or the utilization of machines is low, it could mean the end of the company. Even if the company does not go bankrupt, a wrong schedule can cost a lot of time and money. The goal of my thesis is to develop a webapplication which creates optimal schedules from the data of the products to be manufactured.
Linear programming is one of the most efficient ways, to create schedules. There are numerous linear programming systems available on the market, but they can not create such schedules on their own.
In my thesis I present the typical scheduling problems, and the available linear programming solvers, and then the software I created by merging these. I describe in detail the scheduling problem types, which I implemented my application for. The conversion of the input data is different for each type, so I described them with a great deal of emphasis.
My application recieves the input data on multiple channels, then it converts that data, so the chosen solver can deal with it. Then the result of the solver needs to be converted back to my own model, so that my program can draw the Gantt-chart of the schedule. My model can handle, and work on the datas of every implemented problem type.