SWFS

cpp codes to simulate the production of wafer lots in Semiconductor Wafer Fabrication Systems

Simulation Model

The SWFS is implemented in source file Fab.cpp with header file Fab.h. There are 12 stations each of which may have several identical machines. Meanwhile there is only one type of wafer to produce whose processing procedure consists of 60 steps. You can test the simulation with all default settings by simplily compile and run the project.

If you want to check the details, you first uncomment the following codes in Fab.h:

//std::string message{};//machine, buffer: old state -> new state;
//void update_state_and_record(const size_t& ei);// ei: event index
//void simulation_and_record();
//void write_to_html(std::string fname = "SWFS-state-update-record.html");
//std::string html_buffers(usint bNo1 = 0, usint bNo2 = 0);
//std::string html_machines_SEL(usint eNo = 0, usint mNo1 = 0, usint mNo2 = 0);
//std::string html_NoIbOb(usint si, usint bNo);

then uncomment the corresponding implementations in the source file Fab.cpp:

//void Fab::update_state_and_record(const size_t& ei) {
// ...
//}
//
//void Fab::simulation_and_record() {
// ...
//}
//
//void Fab::write_to_html(string fname) {
// ...
//}
//
//string Fab::html_buffers(usint bNo1, usint bNo2) {
// ...
//}
//
//string Fab::html_machines_SEL(usint eNo, usint mNo1, usint mNo2) {
// ...
//}
//
//string Fab::html_NoIbOb(usint si, usint bNo) {
// ...
//}

After the two steps, we can use the method simulation_and_record() to run the simulation and the intermediate state changes will be written in file SWFS-state-update-record.html which we can explore with any browser. Note that the head.html and foot.html are building blocks for the former html file.

• lessOb_moreIb(const usint& i, const usint& j) and maxIb_minOb(const usint& si): wafer lots scheduling, select the max incoming buffer with the least outgoing buffer, excluding those outgoing buffers with 0 incoming buffer.
• reset(), add_machine(const usint& sNo) and del_machine(const usint& sNo): reset SWFSs.
• rexp(const float& mean): generating exponential random variables with parameter mean.

Marginal Machine Allocation Algorithms

In order to maximize the throughput of SWFSs by allocating proper number of machines into different stations, I implemented a series of Marginal Machine Allocation Algorithms in margin_alloc.cpp with head file margin_alloc.h. The two algorithms are Marginal One Machine Allocation Algorithm and Marginal Two Machines Allocation Algorithm.

• mass_run_margin_one(Fab& fab, usint N_more, usint Rep = 200, usint rep = 1)
• mass_run_margin_two(Fab& fab, usint N_more, usint Rep = 200, usint rep = 1)
• std::vector<usint_float> greedy_search(const usint& Nmore, Fab& fab, const usint& rep = 1): greedy using add_one()
• std::vector<iandj_float> greedy_search2(const usint& Nmore, Fab& fab, const usint& rep = 1): greedy using add_two()
• usint_float add_one(Fab& fab, const usint& rep = 1): each time allocate one machine
• iandj_float add_two(Fab& fab, const usint& rep = 1): each time allocate two machines