About luboxing

luboxing · ‎06-21-2023

Thanks FreelanceReinh! Yes your code is exactly what I need... I was aware the "table" but of only its vector use. Yes I am trying to simulate the matrix and here is the frequency result from your code. As you can see the marginal probabilities are consistent with my first post! Many thanks! Table of y by x y x 1 2 3 4 5 6 7 8 9 10 Total 1 5.44 2.33 0.39 1.54 0.00 1.16 0.00 0.39 0.00 0.00 11.24 2 1.95 4.66 1.15 0.38 1.94 0.00 0.78 0.40 0.00 0.00 11.26 3 2.29 1.55 4.26 0.39 1.94 0.77 0.00 0.00 0.00 0.00 11.21 4 0.78 0.77 1.15 3.12 0.75 2.36 1.58 0.76 0.00 0.00 11.25 5 0.39 0.38 0.78 2.67 3.50 0.40 1.55 1.18 0.38 0.00 11.24 6 0.39 0.39 1.16 1.16 0.76 2.31 1.54 2.34 0.77 0.39 11.21 7 0.00 0.00 0.77 0.38 0.38 2.31 3.50 2.35 1.16 0.39 11.23 8 0.00 0.76 1.15 1.17 1.55 0.79 1.14 3.90 0.39 0.39 11.24 9 0.00 0.00 0.00 0.39 0.39 0.39 0.37 0.00 3.90 1.17 6.61 10 0.00 0.39 0.39 0.00 0.00 0.78 0.78 0.00 0.00 1.17 3.51 Total 112473 11.25 112322 11.23 111930 11.19 112112 11.21 112032 11.20 112597 11.26 112363 11.24 113066 11.31 65931 6.59 35174 3.52 1000000 100.00

luboxing · ‎06-21-2023

Thanks Ballardw! I sampled from your code for 100 times with macro and here is the resulting distribution (percent*100). It is quite different from the original distribution. Table of y by x y x 1 2 3 4 5 6 7 8 9 10 Total 1 1.78 1.68 0.79 0.50 1.98 0.69 0.59 0.69 0.40 0.89 10.00 2 1.09 0.79 1.49 1.68 1.29 0.79 0.59 0.99 0.99 0.30 10.00 3 0.99 0.89 0.99 1.09 0.59 1.19 1.39 0.99 0.99 0.89 10.00 4 1.09 1.49 0.89 0.79 0.99 0.99 1.39 1.29 0.59 0.50 10.00 5 0.99 0.79 1.58 0.89 1.09 0.69 1.09 0.99 1.29 0.59 10.00 6 1.09 1.09 0.79 1.78 0.69 1.39 1.29 0.79 0.89 0.20 10.00 7 1.49 1.39 1.19 0.99 1.29 1.09 0.50 0.79 0.99 0.30 10.00 8 0.99 0.89 1.29 0.99 0.79 1.39 0.89 1.29 0.89 0.59 10.00 9 0.40 0.89 0.79 1.09 0.50 0.69 1.09 1.19 1.58 1.78 10.00 10 0.10 0.10 0.20 0.20 0.79 1.09 1.19 0.99 1.39 3.96 10.00 Total 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 100.00

luboxing · ‎06-21-2023

Koen, Thanks for replying! Since each column and row can only be drawn once, the YouTube solution wouldn't work because it does not exclude the rows and columns already drawn. In this example, I will need draw 10 times and come up with such series as (1,1), (2,2), (3,3), .... (10,10), but the YouTube solution possibly gives (1,1), (2,2), (3, 2) ... which repeats row 2 twice. Instead, I can still use the YouTube code with a modification: in each iteration I reset the cell already drawn to 0 and then recalculate the densities of all other cells (marginal probabilities). This way, with 10 draws I will get the sample. However, this process is still a bit cumbersome. Just wondering if SAS offers something easier. Bo

luboxing · ‎06-21-2023

Hi, I am trying to simulating data with the following joint pmf: X Y 1 2 3 4 5 6 7 8 9 10 Grand Total 1 5.43% 2.33% 0.39% 1.55% 0.00% 1.16% 0.00% 0.39% 0.00% 0.00% 11.24% 2 1.94% 4.65% 1.16% 0.39% 1.94% 0.00% 0.78% 0.39% 0.00% 0.00% 11.24% 3 2.33% 1.55% 4.26% 0.39% 1.94% 0.78% 0.00% 0.00% 0.00% 0.00% 11.24% 4 0.78% 0.78% 1.16% 3.10% 0.78% 2.33% 1.55% 0.78% 0.00% 0.00% 11.24% 5 0.39% 0.39% 0.78% 2.71% 3.49% 0.39% 1.55% 1.16% 0.39% 0.00% 11.24% 6 0.39% 0.39% 1.16% 1.16% 0.78% 2.33% 1.55% 2.33% 0.78% 0.39% 11.24% 7 0.00% 0.00% 0.78% 0.39% 0.39% 2.33% 3.49% 2.33% 1.16% 0.39% 11.24% 8 0.00% 0.78% 1.16% 1.16% 1.55% 0.78% 1.16% 3.88% 0.39% 0.39% 11.24% 9 0.00% 0.00% 0.00% 0.39% 0.39% 0.39% 0.39% 0.00% 3.88% 1.16% 6.59% 10 0.00% 0.39% 0.39% 0.00% 0.00% 0.78% 0.78% 0.00% 0.00% 1.16% 3.49% Grand Total 11.24% 11.24% 11.24% 11.24% 11.24% 11.24% 11.24% 11.24% 6.59% 3.49% 100.00% One sample of distribution is: X Y 1 2 3 4 5 6 7 8 9 10 Grand Total 1 1 1 2 1 1 3 1 1 4 1 1 5 1 1 6 1 1 7 1 1 8 1 1 9 1 1 10 1 1 Grand Total 1 1 1 1 1 1 1 1 1 1 10 Notice that each checked cell (1) occupies a distinct row and column. Is there a SAS proc or IML functions that simulates from this distribution? Thanks in advance!

luboxing · ‎08-10-2022

Thanks for looking into it, Ballardw! Sure -- for example, if I pick 1/14/2022 as begining day for January, implying 1/13/2022 as ending day MWE = 2875.82 2/12/2022 as begining day for February, implying 2/11/2022 as ending day MWB = 2849.19 and MWE = 3120.61 3/9/2022 as begining day for March, implying 3/8/2022 as ending day MWB = 3123.72 So the total distance is (2849.19-2875.82)**2 + (3123.72-3120.61)**2 = 718.829 and is the optima (min).

luboxing · ‎08-10-2022

Hi, I came across this problem with table below. Here Y, M are year and month, and Y2 and M2 are the next month of (Y, M). In each month here are days (totaling 28/29/30/31 depending on month) designated by DY. For each month, I am going to pick one day as the beginning day of the month (DY) and another day as the ending day (DY1) so that DY is rotated. For example, if 5 is picked as the beginning day, 4 would be the ending day, as days follow 5, 6,..., 31, 1, 2, 3, 4 assuming the month has 31 days, and so on. Once the beginning day has been picked for the month (ending day implied), I move to the next month and do the same. At the end, I need to minimize: sum((MWB(i) - MWE(i+1))**2), where i = 1... # of months. The example below has 3 months so the total space is 31*30*31. The actual problem involves one year and a much larger space of 31*30*31...*31. I am considering optmodel but still trying to specify the problem with it. Any help is appreciated! Y M M2 Y2 DY DY1 MWB MWE 2022 1 2 2022 1 31 3,060.79 2,847.60 2022 1 2 2022 2 1 2,795.53 2,755.11 2022 1 2 2022 3 2 2,711.12 2,723.58 2022 1 2 2022 4 3 2,585.14 2,357.31 2022 1 2 2022 5 4 2,856.69 2,336.30 2022 1 2 2022 6 5 2,585.89 2,314.10 2022 1 2 2022 7 6 2,708.57 2,537.95 2022 1 2 2022 8 7 2,761.23 2,327.29 2022 1 2 2022 9 8 2,550.63 2,352.52 2022 1 2 2022 10 9 2,970.52 2,354.57 2022 1 2 2022 11 10 2,995.84 2,281.79 2022 1 2 2022 12 11 3,453.13 2,612.58 2022 1 2 2022 13 12 3,000.57 2,619.14 2022 1 2 2022 14 13 2,886.60 2,875.82 2022 1 2 2022 15 14 3,054.25 2,646.22 2022 1 2 2022 16 15 2,983.17 2,611.71 2022 1 2 2022 17 16 2,516.30 2,709.25 2022 1 2 2022 18 17 2,670.58 2,623.86 2022 1 2 2022 19 18 2,784.30 2,254.84 2022 1 2 2022 20 19 2,570.44 2,331.93 2022 1 2 2022 21 20 2,451.75 2,354.15 2022 1 2 2022 22 21 2,423.53 2,283.58 2022 1 2 2022 23 22 2,455.70 2,130.18 2022 1 2 2022 24 23 2,512.75 2,028.61 2022 1 2 2022 25 24 2,493.90 2,104.30 2022 1 2 2022 26 25 2,529.80 2,213.52 2022 1 2 2022 27 26 2,574.74 2,228.63 2022 1 2 2022 28 27 3,012.91 2,277.01 2022 1 2 2022 29 28 3,163.53 2,289.61 2022 1 2 2022 30 29 3,372.30 2,690.60 2022 1 2 2022 31 30 3,150.71 2,840.38 2022 2 3 2022 1 30 4,147.38 2,937.49 2022 2 3 2022 2 1 4,774.03 3,289.59 2022 2 3 2022 3 2 4,144.49 4,371.40 2022 2 3 2022 4 3 4,090.07 3,824.04 2022 2 3 2022 5 4 4,085.97 3,807.31 2022 2 3 2022 6 5 3,766.41 3,786.18 2022 2 3 2022 7 6 3,382.19 3,770.40 2022 2 3 2022 8 7 3,331.66 3,463.21 2022 2 3 2022 9 8 2,985.52 3,040.08 2022 2 3 2022 10 9 3,420.95 2,888.73 2022 2 3 2022 11 10 3,336.96 2,751.84 2022 2 3 2022 12 11 2,849.19 3,120.61 2022 2 3 2022 13 12 2,923.18 2,870.00 2022 2 3 2022 14 13 3,466.61 2,598.92 2022 2 3 2022 15 14 3,581.33 2,630.91 2022 2 3 2022 16 15 3,751.99 3,145.85 2022 2 3 2022 17 16 3,793.83 3,315.29 2022 2 3 2022 18 17 3,927.19 3,409.43 2022 2 3 2022 19 18 3,958.74 3,521.23 2022 2 3 2022 20 19 3,767.45 3,619.12 2022 2 3 2022 21 20 3,569.70 3,727.37 2022 2 3 2022 22 21 3,951.63 3,494.98 2022 2 3 2022 23 22 3,893.76 3,311.69 2022 2 3 2022 24 23 3,710.34 3,664.18 2022 2 3 2022 25 24 3,100.67 3,606.88 2022 2 3 2022 26 25 3,164.81 3,345.92 2022 2 3 2022 27 26 2,974.90 2,768.13 2022 2 3 2022 28 27 3,391.46 2,798.68 2022 2 3 2022 29 28 3,377.84 2,738.24 2022 2 3 2022 30 29 3,499.97 3,046.27 2022 3 4 2022 1 31 3,653.04 3,733.46 2022 3 4 2022 2 1 3,992.26 3,705.17 2022 3 4 2022 3 2 4,419.02 3,384.90 2022 3 4 2022 4 3 4,390.42 3,612.65 2022 3 4 2022 5 4 4,330.79 3,914.40 2022 3 4 2022 6 5 3,968.64 3,965.32 2022 3 4 2022 7 6 3,781.22 3,805.89 2022 3 4 2022 8 7 3,256.24 3,613.85 2022 3 4 2022 9 8 3,123.72 3,506.88 2022 3 4 2022 10 9 3,494.57 2,894.91 2022 3 4 2022 11 10 3,665.44 2,818.28 2022 3 4 2022 12 11 3,946.23 3,105.49 2022 3 4 2022 13 12 3,801.13 3,436.77 2022 3 4 2022 14 13 4,145.82 3,608.90 2022 3 4 2022 15 14 4,193.16 3,579.46 2022 3 4 2022 16 15 3,660.22 3,740.13 2022 3 4 2022 17 16 3,764.65 3,778.65 2022 3 4 2022 18 17 3,819.29 3,409.57 2022 3 4 2022 19 18 3,546.83 3,455.49 2022 3 4 2022 20 19 4,045.83 3,596.81 2022 3 4 2022 21 20 4,197.98 3,244.33 2022 3 4 2022 22 21 3,476.33 3,635.28 2022 3 4 2022 23 22 3,451.32 3,780.78 2022 3 4 2022 24 23 3,513.11 3,002.59 2022 3 4 2022 25 24 3,769.57 3,064.38 2022 3 4 2022 26 25 3,588.74 3,116.02 2022 3 4 2022 27 26 4,063.67 3,503.90 2022 3 4 2022 28 27 4,319.61 3,308.17 2022 3 4 2022 29 28 4,127.35 3,703.07 2022 3 4 2022 30 29 4,143.08 3,796.24 2022 3 4 2022 31 30 4,084.32 3,711.78

luboxing · ‎08-10-2022

Thanks, Rob! It works like a charm! My actual usage has i=0..30, j=1-30, and t=1..30. With your code it takes around 15 seconds to solve.

luboxing · ‎08-09-2022

Thanks, Rob! I did encounter one error, however: 57 con FlowBalance {<i,t> in NODES}: 58 sum {<(i),j,(t)> in ARCS} UseArc[i,j,t] - sum {<j,(i),(t)-1> in ARCS} UseArc[j,i,t-1] 59 = (if <i,t> = <SOURCE,0> then 1 else if <i,t> = <SINK,n+1> then -1 else 0); _ 22 76 ERROR 22-322: Syntax error, expecting one of the following: IN, NOT, ~. ERROR 76-322: Syntax error, statement will be ignored. Any ideas?

luboxing · ‎08-08-2022

Thanks for your quick reply, Rob! Could you help me with the syntax? I am a beginner with Proc optmodel and here is what I have so far. PROC OPTMODEL; SET TS = 1..4; SET IDX = 1..4; NUM SOURCE = 0; NUM SINK = 1 + MAX {I in IDX} I; SET NODES = IDX UNION {SOURCE, SINK}; SET <NUM, NUM> ARCS INIT {}; NUM INCS {IDX, IDX}; NUM INCAVG {TS}; READ DATA DYINC_M INTO [I=FRIDX]{J IN IDX} <INCS[I,J]=COL('INC'||J)>; READ DATA HISINCAVG_M INTO [DYIDX] INCAVG=INC; VAR USEARC {ARCS} BINARY; NUM COST {I IN IDX, J IN IDX, T IN TS} = (INCS[I,J] - INCAVG[T])**2; MIN TOTALCOST = SUM{I IN IDX, J IN IDX} COST[I,J]*USEARC[I,J]; ... QUIT;

luboxing · ‎08-08-2022

Greetings! I am faced with a slightly modified TSP problem as below: There are 4 nodes with directional values: FRIDX TOIDX INC 1 2 -0.23548 1 3 -0.14027 1 4 -0.22685 2 1 0.014942 2 3 -0.02719 2 4 -0.11378 3 1 0.076665 3 2 -0.06068 3 4 -0.05206 4 1 0.059285 4 2 -0.07806 4 3 0.017148 However INC above is not the cost in a TSP problem. TO calculate the costs, here is another table: Trip # HIST_INC 1 -0.072240154 2 -0.075171085 3 -0.077492213 Where trip # represent the subtrip number traveling to all nodes above without return. That is, 4 -->2 --> 1 --> 3 requires 3 subtrips. HIST_INC is a historical value for INC I want to align with by defining the cost as: cost = sum (sq(INC - HIST)) for all 3 trips (across 4 nodes). Therefore, the costs are not only determined by the starting and ending nodes and their directions, but also by the order of each subtrip within the entire iteration acorss all nodes. For example, the cost of subtrip from 4 to 1 could take any of the three values, depending on the order of the subtrip within the enture trip (without return): sq(0.059285 + 0.072240154) sq(0.059285 + 0.075171085) sq(0.059285 + 0.077492213) I hope my explanation is clear; if not I'd be glad to further clarify. I looked up PROC OPTNETWORK used it for a standard TSP, but no luck with this one. Thanks in advance! Bo

luboxing · ‎08-08-2022

Thanks, Rob!

luboxing · ‎08-08-2022

Greeating! I am wondering how to specify the problem of the traveling salesman problem (TSP) with different starting and ending nodes. I am able to replicate the proc optnetwork example below described in the SAS manual: proc optnetwork direction = directed logLevel = moderate links = mycas.LinkSetIn outNodes = mycas.NodeSetOut; tsp out = mycas.TSPTour; run; But the example only applies to a TSP problem where the salesman goes back to the same node he started from. The problem I want to solve relaxes that requirement -- he only need to go to every node without returning to the starting point. Any help would be highly appreciated! Bo

luboxing · ‎03-04-2021

Good idea... Just posted!

luboxing · ‎03-04-2021

Hi -- I am trying to generate auto-correlated time-series of say 24 intervals (24 hours in a day) with each interval following a Weibull distribution (of different parameters across hours). My understanding is that each value depends on the last one so they have to be generated in sequence. The training data is thousands of consequtive days (with 24-hours each) from history. I'd also like to differentiate month and weekday(0)/weekend(1). Any thoughts? Thanks, Bo

luboxing · ‎03-03-2021

Rick, I am trying to apply the solution here to a slightly different problem: to generated auto-correlated time-series of say 24 intervals (24 hours in a day) with each interval following a Weibull distribution (of different parameters across hours). My understanding is that each value depends on the last one so they have to be generated in sequence. Any thoughts? Thanks, Bo

Online Status	Offline
Date Last Visited	‎06-22-2023 06:53 PM

Re: Simulating from joint pmf

Re: Simulating from joint pmf

Re: Simulating from joint pmf

Simulating from joint pmf

Re: Least distance day-matching

Least distance day-matching

Re: Modified traveling salesman problem

Re: Modified traveling salesman problem

Re: Modified traveling salesman problem

Modified traveling salesman problem

Re: Solving for optimal scalers

Re: Solving for optimal scalers

Re: Run multiple Process Flows at once using SAS Code (on EG 7.11 HF3)

Re: Simulating from joint pmf

Re: Simulating from joint pmf

Re: Simulating from joint pmf

Simulating from joint pmf

Re: Least distance day-matching

Least distance day-matching

Re: Modified traveling salesman problem

Re: Modified traveling salesman problem

Re: Modified traveling salesman problem

Modified traveling salesman problem

Re: Traveling salesman with different begining and ending nodes

Traveling salesman with different begining and ending nodes

Re: Generate correlated random variables that follow beta distribution...

Generate autocorrelated timeseries with each observation following Wei...

Re: Generate correlated random variables that follow beta distribution...