N::meprec in LinearProgramming
- To: mathgroup at smc.vnet.net
- Subject: [mg84860] N::meprec in LinearProgramming
- From: Yaroslav Bulatov <yaroslavvb at gmail.com>
- Date: Wed, 16 Jan 2008 23:07:19 -0500 (EST)
I'm getting mecprec messages using Linear Programming, even if I use
really large $MaxExtraPrecision, however, the expressions that cause
the error evaluate just fine using N directly,
m = {{5/(2 Sqrt[133]), -(11/(14 Sqrt[57]))}, {-(8/Sqrt[133]), 10/(
7 Sqrt[57])}, {15/(2 Sqrt[133]), 5/(
14 Sqrt[57])}, {-(5/(2 Sqrt[133])), -(65/(14 Sqrt[57]))}, {0, (
2 Sqrt[19/3])/7}, {1/(2 Sqrt[133]), -(25/(14 Sqrt[57]))}};
b = {31/20, -(31/5), 57/10, -(41/20), 0, 0};
LinearProgramming[{1, 0}, m, b, -Infinity]
(* The first expression causing the error is below, and works just
fine with N. *)
N[-((63 Sqrt[3/19])/100) + (287 Sqrt[57])/650 -
1/13 Sqrt[21] (-((3 Sqrt[7/19])/100) + (19 Sqrt[133])/25)]
Why do I get the message in LinearProgramming but not in N? Does the
message imply the solution returned by LinearProgramming could be
incorrect?