Model generation time large compared to solving time

[Simoning]

Hello,

I am working with a relatively large LP model. Currently it takes around 30 minutes to generate the model, although it only takes less than 2 minutes to solve it with Cplex using LP. In the profile summary I find two equations that take very long time to generate, as these blocks consist of many single equations. My question is, do I need to simplify my equations to reduce generation time, or is there any options/settings that I could try? If I need to simplify the model, is there any aspects that are especially important in relation to generation time? (number of decimals, length of variable names, avoiding t--1 notation, etc.)

Best regards, Simon

From the lst. file:
MODEL STATISTICS

BLOCKS OF EQUATIONS 25 SINGLE EQUATIONS 1,095,002
BLOCKS OF VARIABLES 17 SINGLE VARIABLES 1,384,081 17,520 projected
NON ZERO ELEMENTS 4,248,601


---- 575 Solve Fini PMM 1.407 1723.641 SECS 622 MB 4248601
GENERATION TIME = 1723.250 SECONDS 622 MB 38.2.1 96226ea8 WEX-WEI
---- 575 Solve Alg PMM 99.703 1823.344 SECS 622 MB

---- Profile Summary (41 records processed)
1174.140 0.484GB 488 Equation energy_use (210240)
441.922 0.493GB 493 Equation energy_prod (78840)
99.703 0.622GB 575 Solve Alg PMM
66.046 0.608GB 531 Equation mb_feedwater (8760)
17.500 0.521GB 517 Equation mb_steam (35040)
10.328 0.243GB 362 Equation obj_cost (1)
9.640 0.622GB 590 Execute
8.094 0.622GB 592 Execute
7.969 0.622GB 591 Execute
7.813 0.506GB 501 Equation flow_steam_turb (87600)

[Fred]

Hi Simon,

Half an hour model generation time for a model of that size is indeed incredibly long. Hard to say why it takes so long without being able to reproduce.

Here we have some hints on how to implement models efficiently.

If you could share the model/data, it might be easier for other forum users to provide target-oriented help.

Best,
Fred