HP50G SIMPLEX Version14b Max Min Pivot Algorithm, multiple/unbounded, unfeasible sol - Printable Version +- HP Forums (https://www.hpmuseum.org/forum) +-- Forum: HP Software Libraries (/forum-10.html) +--- Forum: General Software Library (/forum-13.html) +--- Thread: HP50G SIMPLEX Version14b Max Min Pivot Algorithm, multiple/unbounded, unfeasible sol (/thread-19215.html) Pages: 1 2 |
RE: HP49-50G VERSION 12.hp SIMPLEX Max Min Pivot Algorithm, multiple/unbounded solutions - Gil - 01-26-2023 03:20 PM Version 12 - In some very special cases, the program could be sometimes "stuck" at the end and chose the wrong row, returning then the second solution — but possibly an unfeasible one. A special "artificial" test was duly included in subprogram CHECK. 2SOL to fix up that issue. - The NOTE Var was changed : explanations now only regarding the "pivoting" (outside the program) & no more explanations relative to roundings. - Lines of explanations in —>PIVOT¦&COL.ROW subprogram suppressed for speed purpose. RE: HP49-50G VERSION 12.hp SIMPLEX Max Min Pivot Algorithm, multiple/unbounded solutions - Gil - 01-26-2023 03:28 PM New Version 12 - In some very special cases, the program could be sometimes "stuck" at the end and chose the wrong row, returning then the second solution — but possibly an unfeasible one. A special "artificial" test was duly included in subprogram CHECK.2SOL to fix up that issue. - The NOTE Var was changed : explanations now only regarding the "pivoting" (outside the program) & no more explanations relative to roundings. - Lines of explanations in —>PIVOT¦&COL.ROW subprogram suppressed for speed purpose. RE: HP49-50G VERSION 13d.hp SIMPLEX Max Min Pivot Algorithm, multiple/unbounded solutions - Gil - 01-29-2023 02:54 PM Version 13d Handling of "unfeasible" problem. For visual outlook, deletion of no more used SAME.SOL. Just look at the following problem (input Matrix) : [[ 1 1 'L' 5 ]* [ 0 1 'G' 8 ]** [ 6 4 0 'Max' ]] According to usual SIMPLEX convention (in this program), if nothing is said about xi (here x1 and x2), it is assumed to be >=0. If x2 >= 8 according to **, then * cannot be <0 as said in *, unless breaking the above rule and accepting that x1 might be <= 0 (x1=-3). Put [[ 1 1 'L' 5] [ 0 1 'G' 8 ] [ 6 4 0 'Max' ]] on the stack and launch —>GO. You get then, with the new version 13d: { Var 2¦Steps 4 Time1¦s : Z-Max: 14 :x1: -3 :x2: 8 : Sign { x1 } not allowed: UNFEASIBLE.problem}. Then, pressing afterwards the variable S.1-N, you get also the reminder: :Unfeasible problem Sign incompatibility of { x1 }, [x1 x2... ]: [ -3 8 ] Of course, if you specify that x1 is <= 0 [[ 1 1 'L' 5 ] [ 1 0 'L' 0 ] [ 0 1 'G' 8 ] [ 6 4 0 'Max' ]] and launch —>GO with the new above Matrix, you will get: { Var 2¦Steps 5 Time1¦s : Z-Max: 14 :x1: -3 :x2: 8 }, with no special message. The same applies to [[ 1 1 'L' 5 ] [ 1 0 'F' 0 ] [ 0 1 'G' 8 ] [ 6 4 0 'Max' ]], supposing x1 Free (of unknown sign). RE: HP50G SIMPLEX Version14 Max Min Pivot Algorithm, multiple/unbounded, unfeasible sol - Gil - 02-01-2023 02:08 PM Version 14 - Linear Combination of xi in prior versions supposed all xi to be >= 0 (default case), or free. Now, in the linear combination of the xi, besides the free case of xi being free, the default case respects the real signs of the xi. - Programs variables S.1N, S.LIN & d.LIN, when called, give more details about the solution (multiple/unbounded/unfeasible solution). RE: HP50G SIMPLEX Version14 Max Min Pivot Algorithm, multiple/unbounded, unfeasible sol - Gil - 02-07-2023 09:02 AM New Version 14b Slight change in output for RESULT¦Ci.rows in case of unfeasible problem. |