MATLAB: Linprog() with many constraints

constraintslinprogMATLABoptimiseoptimize

Hi,

I'm trying to optimise the energy mixture in a power system. I've the following problem:

____________________________________

prob = optimproblem('Objective',(Inv_w/LT_w*Pcap_w+Fix_w*(Pcap_w+18)+sum(Var_w*(Pcap_w+18)*Ppu_w)+ Inv_s/LT_s*Pcap_s+Fix_s*Pcap_s+sum(Var_s*Pcap_s*Ppu_s)+Inv_t/LT_t*Pcap_t+Fix_t*Pcap_t+sum(Var_t*Pcap_t*Ppu_t)),'ObjectiveSense','min');

____________________________________

The production has to meet the demand at any hour throughout the year. I know I can set one constraint for each hour manually, but I doubt this is the most efficient use of my time:

____________________________________

prob.Constraints.c1 = (Pcap_w+Kw)*Ppu_w(1)+Pcap_s*Ppu_s(1)+Pcap_t*Ppu_t(1) >= demand(1);

prob.Constraints.c2 = (Pcap_w+Kw)*Ppu_w(2)+Pcap_s*Ppu_s(2)+Pcap_t*Ppu_t(2) >= demand(2);

.

.

.

prob.Constraints.c8760 = (Pcap_w+Kw)*Ppu_w(8760)+Pcap_s*Ppu_s(8760)+Pcap_t*Ppu_t(8760) >= demand(8760);

____________________________________

Is there any way I can set the constraints for each hour throughout a loop or something?

Thanks on advance!

Best Answer

Did you try

constr = (Pcap_w+Kw)*Ppu_w + Pcap_s*Ppu_s + Pcap_t*Ppu_t >= demand;
prob.Constraints.constr = constr;

Alan Weiss

MATLAB mathematical toolbox documentation

Related Solutions

MATLAB: Is there a difference between imported and exported ARXML file with Internal-Constrs

The issue seems to be that the physical range is not taken into account for that type along with the compu method. The type uses a compu method with an internal-to-physical conversion rule of 1/100, which maps on to an internal constraint of 0 to 32000. ceil(log2(32000-0)) is 15, implying we need 15 bits to represent the physical range of 0 to 320, not ceil(log2(320-0)) = 9, which is what we are importing as a numeric type and exporting to the ARXML (the physical constraint of 0 to 5.11 is thus re-calculated from the new internal constraint of 0 to 2^9-1, using the internal-to-physical conversion rule of 1/100).

If you add an internal constraint to "AB10477DC552B6F51113954083CCBE70" that sort of mirrors the physical constraint once the compu method is applied, e.g.

Imported ARXML file: (the addition of the constraint is in bold. It was previously unspecified, but still appeared in the exported ARXML file)

 <DATA-CONSTR UUID="54237557-8678-44b5-b265-189b2e402545">
 <SHORT-NAME>AB10477DC552B6F51113954083CCBE70</SHORT-NAME>
 <DATA-CONSTR-RULES>
 <DATA-CONSTR-RULE>
 <PHYS-CONSTRS>
 <LOWER-LIMIT INTERVAL-TYPE="CLOSED">0</LOWER-LIMIT>
 <UPPER-LIMIT INTERVAL-TYPE="CLOSED">320</UPPER-LIMIT>
 <UNIT-REF DEST="UNIT">/DataType/DataTypeUnits/KiloMtrPerHr</UNIT-REF>
 </PHYS-CONSTRS>
 *<INTERNAL-CONSTRS>
 <LOWER-LIMIT INTERVAL-TYPE="CLOSED">0</LOWER-LIMIT>
 <UPPER-LIMIT INTERVAL-TYPE="CLOSED">32000</UPPER-LIMIT>
 </INTERNAL-CONSTRS>*
 </DATA-CONSTR-RULE>
 </DATA-CONSTR-RULES>
 </DATA-CONSTR>

and we export this to:

Exported ARXML file:

 <DATA-CONSTR UUID="54237557-8678-44b5-b265-189b2e402545">
 <SHORT-NAME>AB10477DC552B6F51113954083CCBE70</SHORT-NAME>
 <DATA-CONSTR-RULES>
 <DATA-CONSTR-RULE>
 <PHYS-CONSTRS>
 <LOWER-LIMIT INTERVAL-TYPE="CLOSED">0</LOWER-LIMIT>
 <UPPER-LIMIT INTERVAL-TYPE="CLOSED">320</UPPER-LIMIT>
 <UNIT-REF DEST="UNIT">/DataType/DataTypeUnits/KiloMtrPerHr</UNIT-REF>
 </PHYS-CONSTRS>
 <INTERNAL-CONSTRS>
 <LOWER-LIMIT INTERVAL-TYPE="CLOSED">0</LOWER-LIMIT>
 <UPPER-LIMIT INTERVAL-TYPE="CLOSED">32767</UPPER-LIMIT>
 </INTERNAL-CONSTRS>
 </DATA-CONSTR-RULE>
 </DATA-CONSTR-RULES>

The internal constraint is 32767, which is 2^15-1, since we need 15 bits to represent the type. If you don't perform this change at all on round trip, you should just use 32767 as the upper limit.

MATLAB: Constraint of optimization problem

It's the wrong solver. You would need to use ga() for a problem like that.

Related Question