I'm trying to make a program run as accurately as possible while staying at a fixed frame rate. How do you do this?
Formally, I have some parameter b in [0,1] that I can set to determine how accurate my computations are (where 0 is least accurate, 0.5 is fairly accurate, and 1 is very accurate). The higher this is, the lower frame rate I will get.
However, there is a "lag", where after changing this parameter, the frame rate won't change until d milliseconds afterwards, where d can vary and is unknown.
Is there a way to change this parameter in a way that prevents "wiggling"? The problem is that if I am experiencing a low frame rate, if I increase the parameter then measure again, it will only be slightly higher, so I will need to increase it more, and then the framerate will be too slow, so I need to decrease the parameter, and I get this oscillating behavior. Is there a way to prevent this? I need to be as reactive as possible in doing this, because changing too slowly will cause the framerate to be incorrect for too long.
Looks like you need an adaptive feedback dampener. Trying an electrical circuit analogy :)
I'd first try to get more info about how the circuit's input signal and responsiveness look like. So I'd first make the algorithm update b not with the desired values but with the previous values plus or minus (as needed towards the desired value) a small fixed increment, say .01 instead (ignore the sloppy response time for now). While doing so I'd collect and plot/analyze the "desired" b values, looking for:
the general shape of the changes: smooth or rather "steppy" or "spiky"? (spiky would require a stronger dampening to prevent oscillations, steppy would require a weaker dampening to prevent lagging)
the maximum/typical/minimum changes in values from sample to sample
the distribution of the changes in values from sample to sample (I'd plan the algorithm to react best for changes in a typical range, say 20-80% range and consider acceptable lagging for changes higher than that or oscillations for values lower than that)
The end goal is to be able to obtain parameters for operating alternatively in 2 modes:
a high-speed tracking mode (also the system's initial mode)
a normal tracking mode
In high-speed tracking mode the b value updates can be either:
not dampened - the update value is the full desired value - only if the changes shape is not spiky and only in the 1st b update after entering the high-speed tracking mode. This would help reduce lagging.
dampened - the update delta is just a fraction (dampening factor) of the desired delta and reflects the fact that the effect of the previous b value update might not be completely reflected in the current frame rate due to d. Dampening helps preventing oscillations at the expense of potentially increasing lag (always conflicting requirements). The dampening factor would be higher for a smooth shape and smaller for a spiky shape.
Switching from high-speed tracking mode to normal tracking mode can be done when the delta between b's previous value and its desired value falls below a certain mode change threshold value (eventually maintained for a minimum number of consecutive samples). The mode change threshold value would be initially estimated from the info collected above and/or empirically determined.
In normal tracking mode the delta between b's previous value and its desired value remain below the mode change threshold value and is either ignored (no b update) or and update is made either with the desired value or some average one - tiny course corrections, keeping the frame rate practically constant, no dampening, no lagging, no oscillations.
When in normal tracking mode the delta between b's previous value and its desired value goes above the mode change threshold value the system switches again to the high-speed tracking mode.
I would also try to get a general idea about how the d response time looks like. To do that I'd change the algorithm to only update b with the desired values not at every iteration, but every n iterations apart (maybe even re-try for several n values). This should indicate how many sample periods would generally a b value change take to become fully effective and should be reflected in the dampening factor: the longer it takes for a change to take effect the stronger the dampening should be to prevent oscillations.
Of course, this is just the general idea, a lot of experimental trial/adjustment iterations may be required to reach a satisfactory solution.
Related
I need to monitor an AC Voltage waveform and record the RMS value when the breakdown happens. I roughly know how to acquire data from videos I have watched, however, it is difficult for me to produce a solution that reads the Breakdown Voltage Value. Ideally, I would also take a screenshot along with the breakdown voltage value,
In case you are not familiar with this topic, When a breakdown happens the voltage will drop immediately to zero. So what I need is to measure the voltage just before it falls to zero, and if possible take a screenshot. This is an image of a normal waveform (black) with a breakdown one (red).
Naive solution*:
Take the data and get the Y values (this would depend on the datatype you have, which would depend on how you acquire the data).
Find the breakdown point by iterating over the values and maintaining a couple of flags (I would probably say track "got higher than X" and once that's true, track "got lower than Y").
From that, I would just say take the last N points (Get Array Subset) and get the array max. Or just track the maximum value as you run.
Assuming you have the graph in a control, you can just right click and select Create>>Invoke Node>>Export Image.
I would suggest trying playing with that with a VI with static data which you can repeatedly run to check how your code behaves.
*I don't know the problem domain and an not overly familiar with the various analysis VIs that ship with LV, so there are quite possibly more efficient ways of doing this.
I am trying to find a way to determine when a value is stabilized in excel. In this case, temperature. I have a set of time-series data, TIMESTAMP and TEMPERATURE. at T=0, the temperature begins to rise at a rapid rate. When the setpoint temperature is approached, the rate of rise decreases, until finally, it stabilizes around the setpoint. The temperature tends to overshoot and undershoot the setpoint by several degrees before stabilization.
How can I have excel figure out when the temperature is "Stable" around the setpoint? For example, Delta T (SP-PV) <= 2. (Delta T is difference between Set Point SP and Process Variable PV)
I was thinking possibly using a time variable to determine a data set size (i.e. 5 minutes) and see if the average Delta T within that time is <= 2.
I do not know how to get excel to search through the whole series effectively though.
If you only look at the difference between the setpoint and the value at some time 0, you may mistake a crossing as a "stabilized" point. You need to examine both the estimated rate of change (Tn - Tn-1)/dt and the current value.
Stability is going to be a combination of "how fast is my signal still changing" and "how close am I". It isn't uncommon in real systems for there to be a fixed offset (steady state error) between a setpoint and a measurement with no change happening in the signal over time. This is typically corrected by an integrative term in the control system.
I am working on the power management of a device using Q-learning algorithm. The device has two power modes, i.e., idle and sleep. When the device is asleep, the requests for processing are buffered in a queue. The Q-learning algorithm looks for minimizing a cost function which is a weighted sum of the immediate power consumption and the latency caused by an action.
c(s,a)=lambda*p_avg+(1-lambda)*avg_latency
In each state, the learning algorithm takes an action (executing time-out values) and evaluates the effect of the taken action in next state (using above formula). The actions are taken by executing certain time-out values from a pool of pre-defined time-out values. The parameter lambda in above equation is a power-performance parameter (0_<lambda<1). It defines whether the algorithm should look for power saving (lambda-->1) or should look for minimizing latency (lambda-->0). The latency for each request is calculated as queuing-time + execution-time.
The problem is that the learning algorithm always favors small time-out values in sleep state. It is because the average latency for small time-out values is always lower, and hence their cost is also small. When I change the value of lambda from lower to higher, I don't see any effect in the final output policy. The policy always selects small time-out values as best actions in each state. Instead of average power and average latency for each state, I have tried using overall average power consumption and overall average latency for calculating cost for a state-action pair, but it doesn't help. I also tried using total energy consumption and total latency experinced by all the request for calculating cost in each state-action pair, but it doesn't help either. My question is: what could be a better cost function for this scenario? I update the Q-value as follows:
Q(s,a)=Q(s,a)+alpha*[c(s,a)+gamma*min_a Q(s',a')-Q(s,a)]
Where alpha is a learning rate (decreased slowly) and gamma=0.9 is a discount factor.
To answer the questions posed in the comments:
shall I use the entire power consumption and entire latency for all
the requests to calculate the cost in each state (s,a)?
No. In Q-learning, reward is generally considered an instantaneous signal associated with a single state-action pair. Take a look at Sutton and Barto's page on rewards. As shown the instantaneous reward function (r_t+1) is subscripted by time step - indicating that it is indeed instantaneous. Note that R_t, that expected return, considers the history of rewards (from time t back to t_0). Thus, there is no need for you to explicitly keep track of accumulated latency and power consumption (and doing so is likely to be counter-productive.)
or shall I use the immediate power consumption and average latency
caused by an action a in state s?
Yes. To underscore the statement above, see the definition of an MDP on page 4 here. The relevant bit:
The reward function specifies expected instantaneous reward as a
function of the current state and action
As I indicated in a comment above, problems in which reward is being "lost" or "washed out" might be better solved with a Q(lambda) implementation because temporal credit assignment is performed more effectively. Take a look at Sutton and Barto's chapter on TD(lambda) methods here. You can also find some good examples and implementations here.
Approximate program behavior:
I have a map image with data associated with the map indicated by RGB index. The data has been populated into an MS Access database. I imported the information in the database into my program as an array and sorted them to go in the order I want the program to run.
I want the program to find the nearest pixel that has a different color from the incumbent pixel being compared. (Colors are stored as string attributes of object Pixel)
First question: Should I use integers to represent my colors instead of string? Would this make the comparison function run significantly faster?
In order to find the nearest pixel of different color, the program begins with all 8 adjacent pixels around the incumbent. If a nonMatch is not found, it then continues onto the next "degree", and in this fashion, it spirals out from the incumbent pixel until it hits a nonMatch. When found, the color of the nonMatch is saved as an attribute of incumbent. After I find the nonMatch for each of the Pixels, the data is re-inserted into the database
The program accomplishes what I want in the manner I've written it, but it is very very slow. After 24 hours, I am only about 3% through with execution.
Question Two: Does my program behavior sound about right? Is this algorithm you would use if you had to accomplish this task?
Question Three: Would it be appropriate for me to use threads in order to finish execution of the program faster? How exactly does that work? (I am brand new to threads, but know a little of the syntax)
Question Four: Would it be more "intelligent" for my program to find the nonMatch for each pixel and insert it into the database immediately after finding it? (I'm making a guess that this would be good in multi-threading, because while one record is accessing the database (to insert), another record is accessing the array of pixels (shared global variable in program).
Question Five: If threading is a good idea, I'm guessing I would split the records up into more manageable chunks (i.e. quarters), and have each thread run the same functions for their specified number of records? Am I close at all?
Please let me know if I can clarify or provide code samples, I just figured that this is more of a conceptual topic so do not want to overburden the post.
1.) Yes, integers compare much faster than strings. Additionally the y use much less memory
2.) I would adapt the algorithm in this way:
E.g.: #1: Let's say, for pixel(87,23) you found the nearest nonMatch to be (88,24) at degree=1 - you can immediately invert the relation and record, that the nearest nonMatch to (88,24) is (87,23). On degree=1 you finished 2 pixels with 1 search.
E.g. #2: Let's say, for pixel (17,18) you found the nearest nonMatch to be (17,20) at degree=2. You can immediately record, that all pixels, that border on both (16,19), (17,19) and (18,19) have the nearest noMatch (17,20) at degree=1, and that one of them is the nearest noMatch to (17,20). On degree=2 (or higher), you finished 5 pixels with 1 search.
3.) Using threads is a two-sided sword: You can do searches in parallel, but you need locking if you write to your array. So this depends on how many CPU cores you can throw at the problem. If this is 3 or more, threads will surely speed up the search.
4.) The results from 2.) make it necessary to mark a pixel as "done" in your array, as you might have finished up to 5 pixels with 1 search. I recommend you put those into a queue and use a dedicated thread to write the queue back into the database: MS Access can't handle concurrent updates, so a single database writer thread looks like a good idea.
5.) I recommend you NOT chunk up the array: You will run into problems with pixels on the edges of a chunk having their nearest nonMatch in a different chunk. Instead if you use e.g. 4 Threads, let them run 1.) From NW corner E, then S 2.) From SE Corner W, then N 3.) From NE Corner S, then W 4. From SW Corner N, then E
Yes. Using a integer would make it much faster
You can reuse the work you have done for previous pixel. Eg. If (a,b) is the nearest non-equal pixel of (x,y), it is likely that points around (x,y) might also have (a,b) as the nearest non-equal pixel
You can use different threads to work on different pixels instead of dividing searching for one pixel
IMHO, steps 1&2 should make your program much faster and you might not need multi-threading.
Yes, I'd convert colour strings to Integers for speed, or even Color structures if you intend to display them on the screen.
Don't work directly with the database if you can avoid it. Copy the necessary data out of the database into an array before you start, and copy your results back when you're finished.
I'm getting a bunch of vectors with datapoints for a fixed set of values, in the example below you see an example of a vector with a value per time point
1D:2
2D:
7D:5
1M:6
6M:6.5
But alas not for all the timepoints is a value available. All vectors are stored in a database and with a trigger we calcuate the missing values by interpolation, or possibly a more advanced algorithm. Somehow I want to be able to tell which data points have been calculated and which have been original delivered to us. Of course I can add a flag column to the table with values indicating whether the value was a master value or is calculated, but I'm wondering whether there is a more sophisticated way. We probably don't need to determine on a regular basis, so cpu cycles are not an issue for determining or insertion.
The example above shows some nice looking numbers but in reality it would look more somethin like 3.1415966533.
The database for storage is called oracle 10.
cheers.
Could you deactivate the trigger temporarily?