Advanced Optimization Technology for global wafer fab scheduling

Based on 10 years of academic and industrial research, Flexciton has developed a hybrid optimization model based on mixed-integer linear programming (MILP) aimed at solving the wafer fab scheduling problem.

The problem with current fab scheduling

Semiconductor wafer fabrication is commonly considered as one of the world's most advanced manufacturing systems. The exceptional complexity, marked by re-entrant flows, mixed processing methods and multiple objectives makes production scheduling extremely difficult.

Extremely difficult​

Multiple objectives

Mixed processing of wafers

Hundreds of process steps

Re-entrant flows

Many fabs adopt dispatch systems to help with production scheduling. These can be highly effective for production control within a single tool or toolset tool. However, because dispatching decisions are made locally and based on pre-defined rules, they fail to account for broader factory behavior.


Improving operational processes is key

To maintain high margins, semiconductor fabs must continue to reduce costs. Historically, this has been accomplished by shrinking chips, shifting production to larger wafers and increasing yield. With little potential left in these options, fabs must now focus on improving operational processes to cut cycle times and increase throughput.


Specifically, to reduce costs and improve efficiency, fabs need to switch to smart scheduling that will:

Increase throughput
Reduce cycle times
Eliminate bottlenecks

Flexciton’s revolutionary scheduling technology for wafer fabs

To meet these needs, we developed a hybrid optimization model based on mixed-integer linear programming (MILP). At the core of our approach is a unique combination of mathematical optimization, heuristic search and smart decomposition methods.

The scheduling software "wish list"

These are the top requirements for scheduling solutions in semiconductor facilities:


Capable of solving complex scheduling problems fast, utilizing unique decomposition technology.


Accounts for all constraints in the optimization model to ensure a true representation of all operations in the fab.


Generates a high-quality solution that best maximizes or minimizes the objective function, with a very small optimality gap.

Easy to manage

Optimizes the scheduling problem consistently, regardless of the objective or the fab state.


Can be scaled to multiple fabs, thanks to the easy configuration of constraints and parameters.

Download our
white paper

and learn more about the new approach to wafer fab scheduling