ELOG GENETIS

Daily Update 6/19/20

Name	Today's Update	Plans for Monday
Alex M	Ran the loop with the same individuals in each "generation" to test how decreasing the grid size would affect the effective volume. Using individuals with lengths of 5cm, I startde with a grid spacing of 0.01 cm and kept cutting down by a factor of 2. I'm down to 0.00125 cm right now. Once I find where the effective volumes stop decreasing, I'll do a binary search for what the optimal grid spacing is for that size.	I'll try doing the same thing that I'm doing for small antennas for large ones to see what the optimal grid size is there, but I think ultimately we're going to want to compare to the progrid to see if XF optimizes better. Amy came up with an idea for finding a function for what the grid spacing should be at a given length, so we'll try to compare that to the progrid. Meaning I'll spend Monday looking through more XF scripting while the large antennas run.
Alex P
Eliot
Leo
Evelyn
Ryan

Thu Jun 18 14:33:32 2020

Daily Update 6/18/20

Name	Update for Today	Plans for tomorrow
Alex M	Alex P and I implemented the changes we discussed with Amy to Detector.cc. We're currently running a test to see how the grid spacing affects the gains from XF. We started with small antennas using a grid spacing of 0.01 and we're cutting down by factors of 2 each generation (until we see that the gains stop changing, then we'll do something of a binary search between the last two grid sizes to search for the largest reasonable size). I also played around with the XF scripting more. My goal is to try to be able to edit the scripts we use to do help us check our results, like save images and load models for the antennas. I can currently run our macro script in XF by loading it in with 1 individual (using a generationDNA.csv file I moved to my user space). I can view and play with the model and it saves an image of it when I run the script (though the image doesn't give much detail). Hopefully this will lead to better understanding XF and being able to check our work and use the progrid option.	I'll continue with the grid space test tomorrow morning and try to play with XF a bit, but I haven't gotten to address many of Amy's comments on the paper from last week so I'll probably prioritize that while the loop is running the AraSim jobs.
Alex P	After asking about the change to replacing negative gains with 0 in the lower frequencies, we implemented that and ran a few individuals through that. We started a normal large antenna run with that change but stopped in order to test the affect of grid spacing on small antennas. We ran a .01 grid spacing and then .005 grid spacing on the same individuals and will continue so on until we see no change in the results. Specifically checking the affects at the 300 MHz frequency but also overall results.	If this works we will find the proper grid spacing for an antenna's size and the next step will be to implement a feature to properly scale that with the antennas. If we still encounter the smaller antennas giving an abnormally high vEffective we will probably have to continue looking into what could be causing that. I won't be online most of tomorrow, I might check in at a few points but Alex M will handle running the tests instead.
Eliot
Leo	Did a run with 5 individ and 5 generations. It successful ran through the loop 5 times, but didn't seem to evolve. In particular, "GenerationDNA" wasn't changing throughout the 5 generations.	Tomorrow, we plan to investigate where this error is coming from within the loop. As of now, we aren't sure whether this is part of the loop or GA. Also we are getting an error message about the file "AraOut_ActualBicone" but doesn't seem to stop the loop, so we want to check in about this as well.
Evelyn
Ryan

Wed Jun 17 17:21:16 2020

Daily Update 6/17/20

Name	Update for Today	Plans for tomorrow
Alex M	Worked more with Alex P on AraSim. We've found the cause of the low effective volumes. The function GainToHeight in Report.cc uses another function called GetGain_1D_OutZero in Detector.cc. The problem is that for some small individuals, we get that the output from GetGain_1D_OutZero is negative, which leads to -nan values from GainToHeight because it takes the square root of the negative values from GetGain_1D_OutZero. We aren't quite sure why we're getting those negative values (we know which variables and operations are giving them, but not why they're so different for some small individuals), but we think a solution is to set any negative output from GetGain_1D_OutZero to just be 0, since they're all lower than the lowest frequency in the band of interest anyway.	We can implement the change to fix these effective volumes. I'm still concerned that we're getting high effective volumes from this, when I would have expected the resolution to have decreased effective volumes for all of the small antennas. We should also be testing this with the actual Bicone to see how it changes that effective volume.
Alex P
Eliot
Leo
Evelyn
Ryan	Possibly made progress with the roulette algorithm for paperclips. However I will not know for sure until I fix the syntax errors in the code ( which seems to be largely just me mistyping something up to this point).	fix the errors in roulette.cpp and make sure the output is desired. Fingers crossed.

Tue Jun 16 16:10:52 2020

Daily Update 6/16/20

Name	Update for today	Plans for Tomorrow
Alex M	Continued working to find the AraSim error. We think we're very close to finding where the difference between the two individuals (one with nonzero effective volume, one with 0) is arising. We've pinned it down to a series of if statements around line 952 of Report.cc. We should be able to piece it apart tomorrow.	Continue working on AraSim to find the error we think it in that if statement.
Alex P
Eliot	Tried to run a couple tests of asymmetric bicone. Finished up on the git codecademy. Getting close to being ready for merge.	Test runs, meeting, hopefully develop new plans.
Leo
Evelyn
Ryan

Mon Jun 15 15:52:45 2020

Daily Update 6/15/20

Name	Today's Update	Plans for Tomorrow
Alex M	Worked with Alex P on looking for the problem in AraSim. We looked in Report.cc where Amy suggested there was an inequality involving a variable called "Full_window" being compared to the trigger threshold. We followed this down to another function in Report.cc where Full_window is used and compared fdiode_real between individuals, which was identical. We then tried printing V_Total_forconvlv, which was different between indviduals, yielding either all 0s or -nans in the individual with 0 effective volume. Helped Ryan some more with working on the roulette algorithm in paperclip. We're going to take it step by step and start with getting a hang of the code with print statements and modifications to see how things work and adding in a function for the roulette algorithm that will simply select the individuals. Once we have that working, we can move onto the next step of properly implementing the roulette algorithm in the code in a manner similar to what we have in the bicone algorithm. I made some small edits in the paper, mostly just pulling some citations from the proceeding but I also think I fixed the format of the references so that the numbers look nicer.	Alex P and I will continue looking in AraSim for the cause of the error. I'll also take some more time to contnue making edits to the paper.
Alex P	Made a generation with an indiviudal we know got a zero fitness score and reduced the grid spacing to see if it would have any affect, it might stay zero as we've done this on a generation before and it lowered all the results so the outliers might not be the zeroes but instead the normal scores with the small antenna. Talked with Amy about looking for errors and started testing what gets passed into the myconvlv function and found that the fdiodes are equal between individuals but the V_Total_forconvlv array was different, and the indiviudal that had zero had a lot more "nans" than the one the had a real score.	Continue investigating the source of where the changes in AraSim that allow for the zero fitness scores come from, marking off one variable at a time to find what makes the difference at each step
Eliot
Leo
Evelyn
Ryan	Ran into more issues attempting to write in the genes for the roulette algorithm for paperclips. So, I'm taking a new approach to attempt to write a roulette function into the existing paperclips evolution to see If I can get it to work before I try to implement the paperclips genes into the roulette algorithm.	Continue writing and attempting to run the roulette function inside of paperclips.

Thu Jun 11 15:14:33 2020

Daily Update 6/11/20

Name	Today's Update	Plans for Tomorrow
Alex M	Helped Alex P look through AraSim some more. We wanted to keep trying to nail down whether or not the threshold was passed (the files said no before, but we wanted to try confirming by looking at the actual values). The files were big but from the values we saw none exceeded the threshold for the individuals with 0 effective volume.	Amy has some more notes for the paper that I'll try to work on. I think we can try semi-running the loop to repeatedly pass in the same individuals and check the results with decreasing grid spacings from XF.
Alex P	Worked with Alex M today to continue to comb through AraSim. Did in fact find the proper place where it checks the trigger and started by printing out the values it compares. When we printed all values instead of just the values that passed we got files that were 5-7 GB and didn't even finish, and weren't told much useful information since the failures were just 0 or -nan. Started by trying to work backwards and see where the value in the trigger comes from, but after looking through multiple files trying to trace back there ended up being a lot of possibilities of where it could have come from so we started looking from the top down and to look where the gainfile's values are used since we now know where the ending is it might be easier to find the path to the end from the start now.	Continue to look through AraSim and trace the gainfile's path to affecting the trigger conditions to see what values would make it pass. Also continuing to investigate the grid spacing's affect on the gain file's values might be worthwhile after we know more about how those values matter. I do have a doctor's appointment tomorrow so will have to leave early and won't make it for group meeting.
Eliot	Finished the plotting software to work and plot everything we want. Entire loop now ~should~ run in its entirety with asymmetry. Looking to do a couple of test runs over the next few days. Worked on CodeCademy git lesson.	Continue work on git lesson and try to get a OSC interactive job. Make a few small edits that might be helpful from a user perspective.
Leo
Evelyn
Ryan

Wed Jun 10 15:47:43 2020

Daily Update 6/10/20

Name	Today's Update	Plans for Tomorrow
Alex M	I wasn't available for most of the day, but I jumped onto the meeting to help Alex P continue looking for the problem with the effective volumes. Amy asked us to investigate the phase responses coming out of XF. I posted the plots in the drop box--see the text below this table for more details.	I think we need to try a run of XF with variable grid spacings so that we can put in the same individuals with increasingly small grid spacings and get a clear view of how that affects the gain/phase vs frequency.
Alex P
Eliot
Leo
Evelyn
Ryan	Continued working on the Roulette algorithm all the function definitions, constants and most of the main function are complete aside from parts that pertain to the genes of the paperclip antennas.	If all goes well, I hope to find the gene parameters used in the paperclip tournament algorithm and then implement those into the work I did today and finish the roulette algorithm completely. If I complete that I plan on testing to make sure everything is working properly before merging onto the main development branch.

For the phase vs frequency plots, see this link: https://www.dropbox.com/home/GP_Antennas/Updates/Phase%20vs%20Frequency%20plots .The plots show the average phase vs frequency for all 10 individuals in one generation of a run. The names of the plot files correspond to which run they come from.The runs which were examined were: XF_Data_Test, Grid_Space_Test, and Ryan_test_run3. They were chosen because they provide use with different antenna, grid spacing, and effective volume data.

The plot for XF_Data_Test is named p_vs_f_XFDT.png. The plot for Grid_Space_Test is named p_vs_f_GST.png. The plot for Ryan_test_run3 is named p_vs_f_Rtr3.png.

XF_Data_Test a grid spacing of 0.1 cm. The antennas in generation 0 were large, but it evolved to make small antennas in generation 5, which is what the phase plot comes from. Individuals 7 and 10 had effective volumes of 0, while the remaining effective volumes were large.
Grid_Space_Test used a grid spacing of 0.01 cm. The antennas in generation 0 were small, which is where the phase plot comes from. None of the individuals had 0 effective volume, but overall the effective volumes were lower than in the similarly sized antennas in generation 5 of XF_Data_Test.
Ryan_test_run3 used a grid spacing of 0.1 cm. The antennas in generation 0, which is where the plot is from, were large and all had similar effective volumes to the actual bicone.

There are two things we noticed about these plots. First, they seem very noisy at low frequencies--they don't have much of a nice pattern here (actually, XF_Data_Test shows generally high phases at low frequencies that decrease with frequency up until around 200 MHz). The other detail we noticed is the periodic behavior of the average phase once it settles down beyond low frequencies. We don't know why this would occur. It's also worth noting that the phase diverge seemingly randomly at the last frequency--we assumed this was similar to how the phase is always wild at theta = 0 and theta = 360 in the .uan files.

118

Thu Nov 19 16:42:06 2020

Daily Update 11/19/20

I hopped on with Alex M to catch him up with what has been done and changed within AraSim so far. I looked into using pointers to access arrays outside of their declared scope and it seemed reletively easy to set up. I set up a pointer corresponding to the arrays V_forfft, T_forfft, and vmmhz_filter and set them at the end of stage one in order to be used in stage two. Also made sure to delete the pointers at the end of stage 2 (Very Important). After this and finding the correct scope to put the deletes in, we complied and didn't get an error in Report.cc. Once it compiled I signed off to study for exam and Alex M will now take over and start to run tests to make sure that with the same seed this gives the same results as a base copy of AraSim. After this the next step would be to start implementing a way to save all the variables from stage one and reading them into stage 2.

112

Fri Oct 9 17:37:57 2020

Daily Update 10/9/20

Name	Update	Plans for Monday
Alex M
Alex P, Ryan, and Ben	Got all functions finished and worked on adding them to main. We set up their calls and commented out the unfunctionalized code with a comment explaining what it was and what date it was commented out. Had to define new variables for limits on our mutation function now that our mutation function is uniformly distributed within a range rather than using standard deviations. Also created some more global variables for maximum outer radius and minimum length in order to have them more accessible.	Finish up implementing and defining variables so that it can run the functions properly. Hoping to have it ready to test on monday.
Eliot
Leo
Evelyn
Ethan	.

117

Fri Oct 30 17:30:16 2020

Daily Update 10/30

Name	Progress	Plans
Alex M
Alex P
Ryan	Attempted to make a new mutation function for the algorithm to try and address the concerns about hitting local maximums from Wednesday. Unfortunately, the idea was unsuccessful when I put it through testing. I would like some input from some of the experts before trying something else. Otherwise, the version I had earlier this week has still been very consistent about optimizing the runs outside of sometimes hitting a local max at about 90/100.
Ben
Ethan	Moved AREA onto my own directory with Alex's help. Began fixing issues with it (small win, no more permissions issues!).	Continue fixing issues with AraSim on my user.
Parker
Elliot
Leo
Evelyn

Draft

Mon Nov 30 16:59:39 2020

Daily Update 10/30

Name	Progress	Plans
Alex M
Alex P
Ryan	Attempted to make a new mutation function for the algorithm to try and address the concerns about hitting local maximums from Wednesday. Unfortunately, the idea was unsuccessful when I put it through testing. I would like some input from some of the experts before trying something else. Otherwise, the version I had earlier this week has still been very consistent about optimizing the runs outside of sometimes hitting a local max at about 90/100.
Ben
Ethan	Moved AREA onto my own directory with Alex's help. Began fixing issues with it (small win, no more permissions issues!).	Continue fixing issues with AraSim on my user.
Parker
Elliot
Leo
Evelyn

116

Mon Oct 26 17:59:10 2020

Daily Update 10/26

Name	Progress	Plans
Alex M
Alex P
Ryan	I have the testing loop finished with plotted results now. The program was able to reach optimal results very quickly. It used 100% tournament selection with a cutoff on the outer radius. The algorithm was using an asymmetric algorithm and the ideal bicone it was being compared to was an arbitrarily picked one from an actual symmetric run individual we knew to stay within the outer radius. All individuals for each generation are plotted on this graph. And the fact that these bicones started as asymmetric shows that we can very easily find symmetric answers if they are indeed ideal.
Ben
Ethan	Tried to fix permission issues with AREA, as well as learned about running the loop and listening to Jorge's thoughts on our project.	Possibly move the AREA project onto the project space as it may solve permissions issues. Alex is looking into it.
Parker
Elliot
Leo
Evelyn

115

Mon Oct 19 17:02:12 2020

Daily Update 10/19

Name	Progress	Plans
Alex M
Alex P
Ryan	Alex P. and I put the outer radius constraint into the Asymmetric our version of the algorithm. I have also created the pseudo-fitness function to be able to do some optimization testing that bypasses the time-consuming parts of the run. All I need to do to finish the pseudo tests is to create a loop to run through the generations and plotting procedures.
Ben
Ethan	Tried to fix permission issues, as well as work on running the loop (invalid id)	Continue fixing AREA permission issues with Ben's help.
Parker
Elliot
Leo
Evelyn

Draft

Fri Oct 16 17:32:49 2020

Daily Update 10/16/20

	Update	Plans for Monday
Alex M
Alex P and Ryan	Made new version of our modified GA that is asymetric. We didn't include the options to start with symetetry but just wanted to get it to a compilable state so that we can test how it would run in the loop to make sure it functions properly there. We also have a real run going with our previous symetrical GA edit.	Watch this run over the weekend and keep it going and then eventually test how the asymetric version works
Eliot
Ben
Leo
Evelyn
Ethan	Worked with Ben and Alex M on getting the AREA project working on slurm. It works now, but there are some permissions issues that we still need to fix. Simply put, Ben can run it perfectly right now.	Find and address permissions issues so anyone can run a job.

113

Mon Oct 12 17:34:29 2020

Daily Update 10/12/20

Name	Update	Plans for Week
Alex M
Alex P, Ryan, and Ben	Finished up writing the updated GA and got it compilable and then got some run time segmentation faults but managed to get all of those fixed. But once we got it running we encountered an error where a lot of the generationDNA ended up being 0,0,0. We looked through our functions and confirmed this was only happening with indiviudals developed through crossover and not mutation or reproduction	Get this problem fixed in crossover and continue to test and make sure this works as intended but so far it is able to generate genDNA files and just has that one bug. The bug is also every other individual so we think we might have messed up how crossover generates two individuals from parents so it should hopefully be an easy fix.
Eliot
Leo
Evelyn
Ethan	.

Tue May 26 16:38:26 2020

Name	Today's Update	Plans for Tomorrow
Alex M.	Continued working on fixing bugs with Alex P. I had a run from over the weekend which got through a few generations, but we're suspicious because the effective volumes are very high despite the antenna dimensions being very small. I made a python script that makes gain plots of the individuals and posted them in the dropbox. I'm not sure they tell us very much. We also got some individuals with effective volumes of 0 in the most recent generations.	In the morning I'll be working with Julie, Alex P., Eliot, and Leo on fixing up some of the git stuff on Eliot's personal space. I also think I should start a run with small antennas but also a small grid size to see if this results in more realistic gain patterns that drive the effective volume down (or up for the ones which had 0 effective volume).
Alex P.	Worked with Alex M to try to find bugs, looked through gain pattern plots but were unable to work out what the problem was. Also worked with Julie and others on getting git working so Eliot and Leo can start their work.	Finish up getting Eliot and Leo set up and practice with git to make sure everything is in order and work on finding error. I want to save all the AraOut files so that we can go back and look at the events passed
Eliot
Leo
Evelyn
Ryan

Fri May 29 17:36:04 2020

Name	Today's Update	Plans for Monday
Alex M	I wasn't available in the morning. I met with Amy and Alex P around 1 to talk about how to use the AraSim function for throwing the same neutrinos to search for bugs in the loop. We want to compare three types of antennas we've evolved to the actual ARA bicone: a small antenna with 0 effective volume, a small antenna with a normal effective volume, and a normal antenna with a normal effective volume. I also spoke with Amy and Julie about the paper and worked on getting the paper in the right format, but I still can't get the citations to work.	I won't be available til late Monday morning, but I'll try to talk to Alex P about using the AraSim function and looking for bugs. Julie and I are supposed to meet at 1 with Ben to discuss him joining, and then I think we have a meeting at 1:30.
Alex P
Eliot
Leo
Evelyn
Ryan

Fri Jun 26 16:13:18 2020

Name	Today's Update	Plans for Monday
Alex M	Worked to run the loop with Alex P to test a smaller initial mean and standard deviation of the opening angle. Worked some more on the paper too--I fixed some citations issues (and found some good ones to add) but the references are still spilling into the margins and off the page.	I'm going to look for where the 3D gain plots in XF for small antennas with different cone separations start to converge so that we can figure out what length we want to set as a cut off minimum to stop them from evolving to be so small.
Alex P
Eliot
Leo
Ryan
Evelyn	(Finally) got Code Academy to work on my browser, started on the git tutotial	Get as far into the git tutorial as I can
Ben
Ethan

Mon Jul 6 16:49:21 2020

Name	Update for Today	Plans for Tomorrow
Alex M	Fixed up a few more errors in the new roulette algorithm with the length cut. Started a run to test the loop. Edited and added a few things to the paper.	My goal is to get a few generations done over night to look at in the morning. If it looks like things are working out, we're going to go ahead and start the process of merging (by adding our edits in to the database version fitst, then Eliot and Leo's asymmetric version, then merging on Wednesday hopefully).
Alex P	Worked with Alex M to have our length test run and also helped ben with an issue with his code when submitted as a job	Look at run in the morning and if it is good start merging and continuing to test the database feature on a variable amount of genes
Eliot
Leo
Evelyn
Ryan	Continued refining the Roulette function in Paperclips. Now negative or zero fitness scores cannot be passed along and the roulette can be isolated from algorithm 1 in the code.	For now, I am continuing to have the last two algorithms mutate for the rest of the individuals but am planning to write in the roulette to create new individuals that are mutated from the chosen. Once I complete this, I plan of coding in a way for user input to decide how many individuals are sent to each algorithm for selection and mutation.
Ben
Ethan

Draft

Tue Aug 11 19:56:45 2020