If you give augcoefmatrix() inequalities, maxima does some odd things:
augcoefmatrix([a = 1, b >= 1, c <= 2], [a, b, c]);
It would be sweet if maxima could handle these. In the case that the inequalities are mixed, eg, the system conatins a subset of inequalities whose operator are <= (call it R), and another subset whose operators are >= (call them S) you can multiply the coefficients by of one of the inequality types by -1 to arrive at a canonical form. This would work the same for < and > also.
It should not be possible to mix >= with >, or <= with > etc.
Perhaps not the best explanation, but I hope you know what I am suggesting.
Log in to post a comment.