OPT-ML DSM Solver

We have integrated the NEOS Optimization Server (https://neos-server.org/neos/) to solve our mixed integer quadratic model using the GAMS/DICOPT solver. Based on known DSM applications, we consider the following three desired objectives of a DSM sequencing algorithm.

DSM Sequencing Algorithm

1.
Minimize the sum of all the lower triangular entries (i.e., the total feedback edges in a binary matrix or feedback information in a real-valued DSM),
2.
Create a block upper triangular matrix so that the sum of the lower triangular entries outside the blocks (i.e., total feedback edges or information among the blocks) is minimized, and
3.
Move the higher feedback dependencies as close as possible to the diagonal.

This graphical version solves up to a 10x10 binary DSM matrix. Please select the size and enter binary upper-triangular (i.e., forward arcs) and lower-triangular (i.e., feedback arcs) entries.

Design Structure Matrix (DSM)

Click the following button to see the sequenced DSM:

Graph of the DSM