PROC OPTMODEL help, constraint

[Lucassss]

Hello,

I am writing a optmodel procedure, with the following constraint, the optimal solution is infeasible:

con unit_con{c in bucket, a in AMOUNTS}: unit[c,a] * Move[c,a] >= 50; 

However, without the above constraint, solution is more reasonable.

What I want to do is to tell the model that if a cell has less than 50 units than don't consider this cell.

The full code is here:

proc optmodel;
/* Index Sets */
set bucket;
set AMOUNTS=2000..17000 by 1000;

/* Parameters */
number orig_amount {bucket, AMOUNTS} init 0;
number loss_amount {bucket, AMOUNTS} init 0;
number profit {bucket, AMOUNTS} init 0;
number unit {bucket, AMOUNTS} init 0;

/* Binary decision variables--approve (1) or deny (0) an amount per customer */
var Move {bucket, AMOUNTS} binary;

read data GRP1 nomiss
   into bucket=[BK_Group]
   	    
   {a in AMOUNTS} < orig_amount[BK_Group,a]=col("OR"||a)
					loss_amount[BK_Group,a]=col("LO"||a)
					profit[BK_Group,a]=col("PR"||a)
					unit[BK_Group,a]=col("UN"||a)>;
print unit;
print move;
print profit;
print orig_amount;

/* Set up amount max and profit max objectives */
min Total_delta_loss = sum{c in bucket, a in AMOUNTS} loss_amount[c,a] * Move[c,a];

/* Approve at most one loan amount per group */
con One_Amount_con{c in bucket}: sum{a in AMOUNTS} Move[c,a] <= 1;

con worse_con{i in bucket, j in AMOUNTS, k1 in 1..(10-i), k2 in 1000..(17000-j) by 1000}: Move[i,j] + Move[i+k1, j+k2] <= 1;

/*con unit_con{c in bucket}: sum{a in AMOUNTS} unit[c,a] * Move[c,a] >= 20;*/
impvar Total_delta_orig = sum{c in bucket, a in AMOUNTS} orig_amount[c,a] * Move[c,a];
impvar Total_Profit = (Total_delta_orig - Total_delta_loss) * 0.45924 - Total_delta_loss;



/* profit constraint: loss no more than 200,000 of profit */
con profit_con:  Total_Profit >= -200000;
con unit_con{c in bucket, a in AMOUNTS}: unit[c,a] * 1 >= 50; 

problem Problem1 include
   Move
   Total_delta_loss
   unit_con
   One_Amount_con profit_con worse_con 
   ;

use problem Problem1;
solve;

print Move;

create data solution
		from [BK_Group LO]
			={c in bucket, g in AMOUNTS: Move[c,g]^=0}
			amount=Move;
print Total_Profit;
print Total_delta_orig;
quit;

Any help is appreciated!

Lucas

[RobPratt]

Can you please share the data?

[Lucassss]

This is what print unit; produces

unit
	2000	3000	4000	5000	6000	7000	8000	9000	10000	11000	12000	13000	14000	15000	16000	17000
1	109	165	218	129	169	222	184	146	109	90	241	57	29	22	12	15
2	108	145	223	126	116	170	153	96	96	64	192	40	24	16	9	8
3	65	134	171	113	101	155	128	104	79	44	138	39	13	13	10	1
4	59	123	143	92	99	129	113	82	81	49	123	31	11	10	9	1
5	45	121	152	96	89	118	118	69	61	50	99	24	6	6	10	3
6	52	122	125	79	67	125	75	62	49	37	96	18	15	5	6	1
7	47	104	125	69	68	130	101	65	59	42	88	16	13	10	5	5
8	50	89	125	65	67	119	84	55	54	33	75	20	12	8	4	3
9	50	89	109	49	57	113	72	52	35	31	56	8	7	9	3	0
10	21	51	60	50	31	102	54	38	26	34	35	13	8	6	3	0

and this is the structure of Move:

Move
	2000	3000	4000	5000	6000	7000	8000	9000	10000	11000	12000	13000	14000	15000	16000	17000
1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
2	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
3	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
4	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
5	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
6	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
7	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
8	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
9	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
10	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0

[Lucassss]

And Move is supposed to be a matrix with 1s and 0s

[Lucassss · Accepted solution]

I figured out myself.

The constraint should be this:

con unit_con{c in bucket}: sum{a in AMOUNTS} (unit[c,a] * Move[c,a]) >= sum{a in AMOUNTS}(50 * Move[c,a]);