Hi all, I am currently using Numerical Recipes' Simplex to solve LP problems but I am not entirely satisfied by it.Several open source projects dealing with LP related optimization are available and I haven't been able to determine which one is the best.GLPKLP SolveCoinORI am mostly interested in LP for medium scales problems (~1000 variables) and in MIP.I know that Ilog's CPLEX is by far the best product but we are still not sure if it's worth the time and price (we have to adapt some parts of the code in our Pricing Library and if the license expires we'll have to rewrite all of it).So far I have tested LP Solve which has a nice C# interface and whose performace are superior to Numerical Recipes' Simplex.GLPK's interface seems less friendly, and CoinOR has a lot of algorithms to solve MIP.Have you got any feedback about these 3 that you can share?Thanks in advance.

I have found some benchmarks on this excellent website.For LP, CPLEX >> CoinOR > GLPK > LP_SOLVESame for MIP.

I have used LP_SOLVE and CPLEX for a non trivial problem. Both have matlab add-ins (the latter by Tomlab). The CPLEX routines are way, way, way faster. Speedups from >1 hour to <1 minute were not unusual. That, I guess, is why they charge 10 thousand bucks for Tomlab whereas LP_SOLVE is free. You can get a trial (1week) license for cplex.