I am using z3 in visual studio and I want to do a simple check with the following
x > 3
After I run my code, I only get one result back in my model. Why do I only get one result back and how would I do model completion in C#?
Z3 is a satisfiability solver (the S in SMT), i.e., it checks whether there is 1 assignment that satisfies the formula. It can compute sets of solutions in some special cases and if the problems are set up the right way, but without further details I can't tell whether any of this would apply to your problem.
Of course, after getting one satisfying assignment you can add a new constraint that forces the assignment you just got is excluded and then get a new assignment.
Related
I am using the default linear solver in GLOP (which I believe is dual-simplex) and getting ABNORMAL result.
I'd like to use normal simplex instead (as this post describes - https://github.com/google/or-tools/issues/1868) but I don't know how to change this setting. How do I change the solver? (Using C# by the way)
Laurent gives the correct steps above. Here is the practical code to create these parameters and pass to solver (if anyone is interested):
var Params = new Google.OrTools.LinearSolver.MPSolverParameters();
Params.SetIntegerParam(MPSolverParameters.IntegerParam.LP_ALGORITHM, (int) MPSolverParameters.LpAlgorithmValues.PRIMAL);
Solver.ResultStatus solved = solver.Solve(Params);
You need to change the parameters of the solver.
The class is here: http://google.github.io/or-tools/dotnet/classGoogle_1_1OrTools_1_1LinearSolver_1_1MPSolverParameters.html
You need to change the integer parameter http://google.github.io/or-tools/dotnet/classGoogle_1_1OrTools_1_1LinearSolver_1_1MPSolverParameters.html#a3b8427a9f4368fd831af8f8c61fb823c
The parameter to change is http://google.github.io/or-tools/dotnet/classGoogle_1_1OrTools_1_1LinearSolver_1_1MPSolverParameters.html#a7319655592ea63d50ef2a6645e309784
the value to assign is http://google.github.io/or-tools/dotnet/classGoogle_1_1OrTools_1_1LinearSolver_1_1MPSolverParameters.html#a79b59c0c868544afdaa05d89c8f8541f
As per the current documentation (see the last row of the table here), it actually seems that the default mode is primal simplex, so that is what one would have if they did not set anything. You'd have to explicitly set parameters to make it work with dual simplex.
I am working on a calculation module using C#, and I bumped on this :
double v = 4 / 100;
I know this is a wrong initialization that returns v = 0.0 instead of v = 0.04
The c# rules says I must ensure at least one of the member is a double, like this :
double v = (double) 4 / 100;
double v = 4.0 / 100;
However, I have many many initializations of that kind that involves integer variables operations, and I feel lazy to browse my code line by line to detect such mistakes.
Instead, is it possible to get warned by the compiler about this ?
Alright, after some playing around and what not, I have a solution. I used this article to come to this solution.I use StyleCop, so you'll need to get and install that. Then, you can download my C# project MathematicsAnalyzer.
First off, I did not account for all type conversion mismatches. In fact, I only accommodate one part.
Basically, I check to see if the line contains "double" followed by a space. I do know that could lead to false warnings, because the end of a class could be double or any number of other things, but I'll leave that to you to figure out how to properly isolate the type.
If a match is found, I check to see that it matches this regex:
double[ ][A-Za-z0-9]*[ ]?=(([ ]?[0-9]*d[ ]?/[ ]?[0-9]*;)|[ ]?[0-9]*[ ]?/[ ]?[0-9]*d;)
If it does -not- match this regex, then I add a violation. What this regex will match on is any of the following:
double i=4d / 100;
double i = 4d / 100;
double i = 4 / 100d;
double i = 4/ 100d;
double i = 4 /100d;
double i = 4/100d;
double i=4d / 100;
double i=4 / 100d;
double i=4/100d;
Any of the above will not create a violation. As it is currently written, pretty much if a 'd' isn't used, it'll throw a violation. You'll need to add extra logic to account for the other possible ways of explicitly casting an operand. As I'm writing this, I've just realized that having a 'd' on both operands will most likely throw an exception. Whoops.
And lastly, I could not get StyleCop to display my violation properly. It kept giving me an error about the rule not existing, and even with a second pair of eyes on it, we could not find a solution, so I hacked it. The error shows the name of the rule you were trying to find, so I just put the name of the rule as something descriptive and included the line number in it.
To install the custom rule, build the MathematicalAnalyzer project. Close Visual Studio and copy the DLL into the StyleCop install directory. When you open Visual Studio, you should see the rule in the StyleCop settings. Step 5 and 6 of the article I used shows where to do that.
This only gets one violation at a time throughout the solution, so you'll have to fix the violation it shows, and run StyleCop again to find the next one. There may be a way around that, but I ran out of juice and stopped here.
Enjoy!
This article explains how to set up custom Code Analysis rules that, when you run Code Analysis, can show warnings and what not.
http://blog.tatham.oddie.com.au/2010/01/06/custom-code-analysis-rules-in-vs2010-and-how-to-make-them-run-in-fxcop-and-vs2008-too/
I'm writing a program to process a cube using Microsoft Analysis, and I want
bit values to be shown as True/False whenever I browse the cube (lets say in Excel).
I've managed to show null values as some string using 'Unknown', but I can't figure how to do the same with the actual values.
P.S. I do not want to convert it by using a CASE query in the DSV table creation. I want it to be only at the Dimension level.
Thanks.
You might "not want" to rely on a CASE statement in the DSV, but SSAS doesn't give you any other options to achieve what you want. So hold your nose and start coding...
FWIW - I prefer a SQL View (rather than calculations in a DSV) for more re-use and easier testing.
I am wanting to know if there is a methodology to feed calculated values to a built-in Revit command from inside a C# program, and then possibly (based on results, such as whether this makes an element too short or too long for a known "maximum span" of a particular beam) continue with my C# program and change the beam size). I am told you can invoke the Revit built-in command after execution of your c# external command, but you cannot then return to the c# program
As another example, I want to select an element to trim/extend to, and have the code figure out which "Joist" beams to extend to this element. My program would do extended filtering (such as "Reference Level", or "Workset", or "Comments", or "Mark" parameters (etc.)) and then run the built in function, providing the element to extend to and then each of my beams.
I've tried internet searches, as well as the Revit SDK samples, and nothing obviously used this (but there are a lot of csproj's to look through).
Can anyone verify that you cannot go back and forth between the C# program and the Revit built-in command?
You can programmatically invoke a built in Revit command with the UIApplication.PostCommand() method. Refer to documentation and building coder for more information. It will not execute until after the API context is over, however.
I don't think you'll be able to feed arguments into the command however, short of some kind of Win32 hack. Perhaps you will need to recreate the functionality of the built in command within the Revit API.
Unfortunately, I don't think we can do (command "_line" pnt1 pnt2) type of thing here.
Perhaps start with the SDK sample "MoveLinear". It shows how to modify end points of linear elements (which includes beams).
The main part of the sample's code is
Autodesk.Revit.DB.Line line;
//get start point via "get_EndPoint(0)"
Autodesk.Revit.DB.XYZ newStart = new XYZ(
lineLoc.Curve.GetEndPoint(0).X + 100,
lineLoc.Curve.GetEndPoint(0).Y,
lineLoc.Curve.GetEndPoint(0).Z);
//get end point via "get_EndPoint(1)"
Autodesk.Revit.DB.XYZ newEnd = new XYZ(
lineLoc.Curve.GetEndPoint(1).X,
lineLoc.Curve.GetEndPoint(1).Y + 100,
lineLoc.Curve.GetEndPoint(1).Z);
//get a new line and use it to move current element
//with property "Autodesk.Revit.DB.LocationCurve.Curve"
line = Line.CreateBound(newStart, newEnd);
lineLoc.Curve = line;
Which moves the X of the first point and the Y of the second point 100 feet.
you can try:
RevitCommandId commandId = RevitCommandId.LookupPostableCommandId(PostableCommand.PlaceAComponent);
commandData.Application.PostCommand(commandId);
I'm not sure if SO is the proper place for asking this, if it's not, I will remove the question and try it in some other place. Said that I'm trying to do the following:
I have a List<double> and want to replace the block of values whose values are situated very close (say 0.75 in this example) to a single value, representing the mean of the replaced values.
The values that are isolated, or alone, should not be modified.
Also the replaced block can't be longer than 5.
Computing the mean value for each interval from 0, 5, 10.. would not provide the expected results.
It happened many times that LINQ power surprised me gladly and I would be happy if someone could guide me in the creation of this little method.
What I've thought is to first find the closest values for each one, calculate the distance, if the distance is less than minimum (0.75 in this example) then assign those values to the same block.
When all values are assigned to their blocks run a second loop that replaces each block (with one value, or many values) to its mean.
The problem that I have in this approach is to assign the "block": if several values are together, I need to check if the evaluating value is contained in another block, and if it so, the new value should be in that block too.
I don't know if this is the right way of doing this or I'm over complicating it.
EDIT: the expected result:
Although you see two axes only one is used, the List is 1D, I should have drawn only the X axis.
The length of the lines that are represented is irrelevant. It's just to mark on the axis where the value is situated.
It turns out that MSDN has already done this, and provided an in-depth example application with code:
Data Clustering - Detecting Abnormal Data Using k-Means Clustering