Method for maximizing the throughput of a multiple-step workstation in a plant

A method for maximizing throughput of a multiple-step workstation in a plant, the workstation having a waiting pool buffer of lots waiting to be processed by the workstation, an nth lot L.sub.n in the waiting pool having a process time PT.sub.n and using a recipe R.sub.n, wherein a delay time DT.sub.n is the time the lot L.sub.n is waiting to be processed due to a preceding lot L.sub.n-1 being processed, the lot L.sub.n not being processed or not being in transit between process steps during the delay time DT.sub.n, n being an integer greater than zero, said method comprising:

• forming an array of interrecipe delay times having a plurality of rows of delay times for a plurality of lots L.sub.n with a plurality of recipes R.sub.n caused by a lot L.sub.n-1 with a recipe R.sub.n-1, and a plurality of columns of delay times for the lot L.sub.n with the recipe R.sub.n caused by a plurality of lots L.sub.n-1 with a plurality of recipes R.sub.n-1 ;
• forming an array of feasible recipe-to-recipe sequences having a plurality of rows of feasible recipe-to-recipe sequences for the lot L.sub.n in a plurality of orders O.sub.m, and a plurality of columns of feasible recipe-to-recipe sequences for a plurality of lots L.sub.n in the order O.sub.m, m being an integer greater than zero;
• forming a lot-recipe mapping array that maps a recipe to be used for an associated lot, wherein each mapping is determined from past processing data;
• generating an objective function for determining the recipe-to-recipe sequences having lots with minimized delay times within an order to be processed;
• determining a constraint function on the recipe-to-recipe sequences;
• generating a linear integer programming model for solving the objective function subject to the constraint function;
• and
• forming a solution set of an optimal process of recipe-to-recipe sequences having lots in an order to be processed according to the linear integer programming model, wherein each lot in the optimal process has a minimum delay time for the order to be processed, the optimal process having a minimum total delay time for the orders to be processed.