I want to design logics similar to a counter in Z3py.
If writing python script, we usually define a variable "counter" and then keep incrementing it when necessary. However, in Z3, there is no variant. Therefore, instead of defining an variant, I define a trace of that variant.
This is a sample code. Suppose there is an array "myArray" of size 5, and the elements in the array are 1 or 2. I want to assert a constraint that there must be two '2's in "myArray"
from z3 import *
s = Solver()
myArray = IntVector('myArray',5)
for i in range(5):
s.add(Or(myArray[i]==1,myArray[i]==2))
counterTrace = IntVector('counterTrace',6)
s.add(counterTrace[0]==0)
for i in range(5):
s.add(If(myArray[i]==2,counterTrace[i+1]==counterTrace[i]+1,counterTrace[i+1]==counterTrace[i]))
s.add(counterTrace[5]==2)
print s.check()
print s.model()
My question is that is this an efficient way of implementing the concept of counter in Z3? In my real problem, which is more complicated, this is really inefficient.
You can do this but it is much easier to create the sum over myArray[i] == 2 ? 1 : 0. That way you don't need to assert anything and you are dealing with normal expressions.
Related
I am solving a serie of linear programing problems using linprog by indexing each problem in a for-loop:
from scipy.optimize import linprog
for i in range(1,N):
sol[i] = linprog(coa3[N][0], A_ub = coa4[i], b_ub = chvneg[i], options= {"disp": True})
I would like to save in a list (still indexed over i) the function minimization result and the array displaying the values of the variables. I guess I need to add in the for-loop something like minfval[i] = ??? and xval[i] = ???, but actually I don't know how to extract these values from the results provided by linprog. Any suggestions? Thanks a lot.
Consider reading the docs as it's pretty clearly explained what exactly linprog returns.
The good thing is, that you are actually storing these values with your code already because you are storing the whole scipy.optimize.OptimizeResult.
It's now just a matter of accessing it:
from scipy.optimize import linprog
for i in range(1,N):
sol[i] = linprog(coa3[N][0], A_ub = coa4[i], b_ub = chvneg[i], options= {"disp": True})
# iterate over solution-vectors
for i in range(1,N):
print(sol[i].x)
# iterate over objectives
for i in range(1,N):
print(sol[i].fun)
You should also check out the field success just to be safe!
Extraction of the docs (link above) on what is contained in linprog's result:
x : ndarray
The independent variable vector which optimizes the linear
programming problem.
fun : float
Value of the objective function.
slack : ndarray
The values of the slack variables. Each slack variable corresponds
to an inequality constraint. If the slack is zero, then the
corresponding constraint is active.
success : bool
Returns True if the algorithm succeeded in finding an optimal
solution.
status : int
An integer representing the exit status of the optimization::
0 : Optimization terminated successfully
1 : Iteration limit reached
2 : Problem appears to be infeasible
3 : Problem appears to be unbounded
nit : int
The number of iterations performed.
message : str
A string descriptor of the exit status of the optimization.
I have a set of IDs associated with costs which is just a double value. IDs are integers and unique. Two IDs may have same costs. I stored them as:-
a=containers.Map('KeyType','uint32','ValueType','double');
a(1)=7.3
a(2)=8.4
a(3)=7.3
Now i want to find the minimum cost.
b=[];
c=values(a);
b=[b,c{:}];
cost_min=min(b);
Now i want to find all IDs associated i.e. 1 and 3 with the minimum cost i.e. 7.3. I can collect all the keys into an array and then do a for loop over this array. Is there a better way to do this entire thing in Matlab so that for loops are not required?
sparse matrix can work as a hashmap, just do this:
a= sparse(1:3,1,[7.3 8.4 7.3])
find(a == min(nonzeros(a))
There are methods which can be used on maps for this kind of operations
http://se.mathworks.com/help/matlab/ref/containers.map-class.html
The approach finding minimum values and minimum keys can be done something like this,
a=containers.Map('KeyType','uint32','ValueType','double');
a(1)=7.3;
a(3)=8.4;
a(4)=7.3;
minval = inf;
minkeys = -1;
for k = keys(a)
val = a.values(k);
val = val{1};
if (val < minval(1))
minkeys = k;
minval = val;
elseif (val == minval(1))
minkeys = [minkeys,k];
end
end
disp(minval);
disp(minkeys);
This is not efficient though and value search is clumsy for maps. This is not what they are intended for. Maps is supposed to do efficient key lookup. In case you are going to do a lot of lookups and this is what takes time, then use a map. If you need to do a lot of value searches, I would recommend that you use a matrix (or two arrays) for this instead.
idx = [1;3;4];
val = [7.3,8.3,7.3];
minval = min(val);
minidx = idx(val==minval);
disp(minval);
disp(minidx);
There is also another post with an example where it is shown how a sparse matrix can be used as a hashmap. Let the index become the key. This will take about 3 times the memory as all non-zero elements an ordinary array, but a map uses more memory than an array as well.
In Python we have lru_cache as a function wrapper. Add it to your function and the function will only be evaluated once per different input argument.
Example (from Python docs):
#lru_cache(maxsize=None)
def fib(n):
if n < 2:
return n
return fib(n-1) + fib(n-2)
>>> [fib(n) for n in range(16)]
[0, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, 233, 377, 610]
>>> fib.cache_info()
CacheInfo(hits=28, misses=16, maxsize=None, currsize=16)
I wonder whether a similar thing exists in Matlab? At the moment I am using cache files, like so:
function result = fib(n):
% FIB example like the Python example. Don't implement it like that!
cachefile = ['fib_', n, '.mat'];
try
load(cachefile);
catch e
if n < 2
result = n;
else
result = fib(n-1) + fib(n-2);
end
save(cachefile, 'result');
end
end
The problem I have with doing it this way, is that if I change my function, I need to delete the cachefile.
Is there a way to do this with Matlab realising when I changed the function and the cache has become invalidated?
Since matlab 2017 this is available:
https://nl.mathworks.com/help/matlab/ref/memoizedfunction.html
a = memoized(#sin)
I've created something like this for my own personal use: a CACHE class. (I haven't documented the code yet though.) It appears to be more flexible than Python's lru_cache (I wasn't aware of that, thanks) in that it has several methods for adjusting exactly what gets cached (to save memory) and how the comparisons are made. It could still use some refinement (#Daniel's suggestion to use the containers.Map class is a good one – though it would limit compatibility with old Matlab versions). The code is on GitHub so you're welcome to fork and improve it.
Here is a basic example of how it can be used:
function Output1 = CacheDemo(Input1,Input2)
persistent DEMO_CACHE
if isempty(DEMO_CACHE)
% Initialize cache object on first run
CACHE_SIZE = 10; % Number of input/output patterns to cache
DEMO_CACHE = CACHE(CACHE_SIZE,Input1,Input2);
CACHE_IDX = 1;
else
% Check if input pattern corresponds something stored in cache
% If not, return next available CACHE_IDX
CACHE_IDX = DEMO_CACHE.IN([],Input1,Input2);
if ~isempty(CACHE_IDX) && DEMO_CACHE.OUT(CACHE_IDX) > 0
[~,Output1] = DEMO_CACHE.OUT(CACHE_IDX);
return;
end
end
% Perform computation
Output1 = rand(Input1,Input2);
% Save output to cache CACHE_IDX
DEMO_CACHE.OUT(CACHE_IDX,Output1);
I created this class to cache the results from time-consuming stochastic simulations and have since used it to good effect in a few other places. If there is interest, I might be willing to spend some time documenting the code sooner as opposed to later. It would be nice if there was a way to limit memory use as well (a big consideration in my own applications), but getting the size of arbitrary Matlab datatypes is not trivial. I like your idea of caching to a file, which might be a good idea for larger data. Also, it might be nice to create a "lite" version that does what Python's lru_cache does.
I understand how the GCD works on a trivial example as below:
for(i=1; i<=100; i++)
{
X[2*i+3] = X[2*i] + 50;
}
we first transform it into the following form:
X[a*i + b] and X[c*i + d]
a=2, b=3, c=2, d=0 and GCD(a,c)=2 and (d-b) is -3. Since 2 does not divide -3, no dependence is possible.
But how can we do this GCD test on a doubly nested loop?
For example:
for (i=0; i<10; i++){
for (j=0; j<10; j++){
A[1+2*i + 20*j] = A[2+20*i + 2*j);
}
}
While the subscripts can be delinearized, the GCD test is simple to apply directly. In your example, the subscript pair is [1+2*i + 20*j] and [2+20*i + 2*j], so we're looking for an integer solution to the equation
1 + 2*i + 20*j = 2 + 20*i' + 2*j'
Rearranging, we get
2*i - 20*i' + 20*j - 2*j = 1
Compute the GCD of all the coefficients, 2, -20, 20, and -2, and see if it divides the constant. In this case, the GCD is 2. Since 2 doesn't divide 1, there's no dependence.
The "easy" way to apply GCD in the nested loop case is to apply it only in cases where the arrays themselves are multidemsional; i.e., the original source code uses multiple subscripts rather than already linearized expressions. Of course if you can "back transform" these linearized subscripts then you'll have the equivalent.
Once you've cast the problem as a multidemsional problem then you may simply apply the GCD test "dimension by dimension". If any dimension shows no dependence then you can stop and declare there is no dependence for the entire multidemsional subscripting sequence.
The key of course is that casting as a multidimensional indexing problem gives you the nice property that there's a one-to-one mapping between individual index values and the corresponding index expression tuples. Without this the problem is harder.
This is the approach I took in the ASC Fortran vectorizing compiler back in the 70's and it worked pretty well, particularly used in conjunction with directional subscript analysis for the non disjoint case. The GCD test by itself is really not sufficient, but it does give you a relatively inexpensive way of making an early decision in your analysis in those cases where you then can avoid the more expensive dependence analysis.
If this is not a real question then feel free to close ;)
Not only the compiler can reorder execution (mostly for optimization), most modern processors do so, too. Read more about execution reordering and memory barriers.
The compiler can change the execution order of statements when it sees fit for optimization purposes, and when such changes wouldn't alter the observable behavior of the code.
A very simple example -
int func (int value)
{
int result = value*2;
if (value > 10)
{
return result;
}
else
{
return 0;
}
}
A naive compiler can generate code for this in exactly the sequence shown. First calculate "result" and return it only if the original value is larger than 10 (if it isn't, "result" would be ignored - calculated needlessly).
A sane compiler, though, would see that the calculation of "result" is only needed when "value" is larger than 10, so may easily move the calculation "value*2" inside the first braces and only do it if "value" is actually larger than 10 (needless to mention, the compiler doesn't really look at the C code when optimizing - it works in lower levels).
This is only a simple example. Much more complicated examples can be created. It is very possible that a C function would end up looking almost nothing like its C representation in compiled form, with aggressive enough optimizations.
Many compilers use something called "common subexpression elimination". For example, if you had the following code:
for(int i=0; i<100; i++) {
x += y * i * 15;
}
the compiler would notice that y * 15 is invariant (its value doesn't change). So it would compute y * 15, stick the result in a register and change the loop statement to "x += r0 * i". This is kind of a contrived example, but you often see expressions like this when working with array indexes or any other base + offset type of situation.