Experiments in Flat Glass Cutting Optimization

The flat glass cutting problem is in the class of NP-Complete problems, thus making an exhaustive search, of every pane in every position, impossible. Instead, an approximate solution is formed, using a combination of software techniques, involving preprocessing of the lists to be cut as well as various feedback processes within the cutting algorithms. Due to practical considerations, the methods used are further restricted to operation on a small personal computer, using an hour or less of processing time, so previously posited solutions are not applicable. This experiment will be limited to polynomial timed solutions, and operates on large (up to 500 panes) data sets. Current results yield a daily average efficiency of up to 89%, excellent when compared to the industry yearly average of 85%. ***************************************************************** Experiments in Flat Glass Cutting Optimization is an investigation of algorithms for the positioning of patterns in the cutting of window panes from stock glass sizes (Fig. A). Through the testing and analysis of various cutting algorithms, insight into the general problem was obtained, and possible solutions posited. SAMPLE PATTERN FOR CUTS OF ’A’, ’B’, & ’C’: +-------------------------+-------+--------------++ | C | | || | | | || +-------------------------+ | || +-------------------------+-------+ B || | | || | | || | A | || | | || | +--------------++ | | | +---------------------------------+---------------+