.NET R&D Digest (November, 2022)

In .NET, November was a full of interesting events – release of .NET 7, bunch of conferences and announcements, in-depth analysis of ASP.NET Core benchmarks, a lot of great posts and of course – a new issue of .NET R&D Digest.

This issue includes bits of software development, a few stories, security, performance, diagnostics, C++, git and of course .NET.

Enjoy your reading!

Continue reading “.NET R&D Digest (November, 2022)”

.NET R&D Digest (June, 2022)

Learning is a continues process. There is literally no end of learning high way because there is always something new, which just emerged, or something that has been with us for years but we never heard of it. This is where .NET R&D Digest comes handy.

This issue of .NET R&D Digest includes bits of development stories about programming and life, architecture, networking, benchmarking, .NET and something to watch.

Enjoy your reading!

Continue reading “.NET R&D Digest (June, 2022)”

Finding “routes” of all-pairs shortest paths with Floyd-Warshall algorithm in C#

Introduction


In the previous post we saw how to solve all-pairs shortest path problem using Floyd-Warshall algorithm. We also explored several ways to improve algorithm’s performance by using parallelism and vectorisation.

However, if you think about the results of the Floyd-Warshall algorithm, you will quickly realise the interesting moment – we know the distances (values of shortest paths) between vertexes in a graph but we don’t know the “routes” i.e. we don’t know what vertexes contribute to each shortest path, and we can’t rebuild these “routes” from the results we have.

In this post, I invite you to join me and extend the Floyd-Warshall algorithm. This time, we are going to make sure we can calculate the distance and rebuild the “route”.

Continue reading “Finding “routes” of all-pairs shortest paths with Floyd-Warshall algorithm in C#”

.NET R&D Digest (October, 2021)

November is here, and in world of .NET it means – .NET Conf and a lot of new and interesting stuff to watch and learn. And one of such things (I hope) – is a new issue of .NET R&D Digest!

Today’s issues contains bits of computer science, software development, .NET, tooling and videos to watch and learn something new!

Enjoy your reading!

Continue reading “.NET R&D Digest (October, 2021)”

.NET R&D Digest (April, 2021)

Today is the last day of April which is also turned out to be a Friday, and, what is always nice to have before weekend? – something to read! 😊 So, here is a new issue of .NET R&D Digest, just for that!

This issue contains bits of computer science, operating systems, security, performance, diagnostics, CSS, .NET and interesting tools to explore.

Enjoy and have a great weekend!

Continue reading “.NET R&D Digest (April, 2021)”

Implementing Floyd-Warshall algorithm for solving all-pairs shortest paths problem in C#

Notice

This post is a free form translation (with some clarifications) of my post in Russian published on techrocks.ru. So if you know Russian, you are welcome to read it there. Otherwise, I invite you to join me here and have some fun.

Introduction


All of us solve a “shortest path” problem many times a day. Unintentionally of course. We solve it on our way work or when we browse the Internet or when we arrange our things on a desk.

Sounds a bit… too much? Let’s find out.

Imagine, you have decided to meet with friends, well … let’s say in cafe. First of all, you need to find a route (or path) to cafe, so you start looking for available public transport (if you are on foot) or routes and streets (if you are driving). You are looking for a route from your current location to cafe but not “any” route – the shortest or fastest one.

Here is a one more example, which isn’t so obvious as the previous one. During you way work you decide to take a “short cut” through a side street because well… it is a “short cut” and it is “faster” to go this way. But how did you know this “short cut” is faster? Based on personal experience you solved a “shortest path” problem and selected a route, which goes through the side street.

In both examples, the “shortest” route is determined in either distance or time required to get from one place to another. Traveling examples are very natural applications of graph theory and “shortest path” problem in particular. However, what about the example with arranging things on a desk? In this case the “shortest” can represent a number or complexity of actions you have to perform to get, for example, a sheet of paper: open desk, open folder, take a sheet of paper vs take a sheet of paper right from desk.

All of the above examples represent a problem of finding a shortest path between two vertexes in a graph (a route or path between two places, a number of actions or complexity of getting a sheet of paper from one place or another). This class of shortest path problems is called SSSP (Single Source Shortest Path) and the base algorithm for solving these problems is Dijkstra’s algorithm, which has a O(n2) computational complexity.

But, sometimes we need to find all shortest paths between all vertexes. Consider the following example: you are creating a map for you regular movements between home, work and theatre. In this scenario you will end up with 6 routes: work -> home, home -> work, work -> theatre, theatre -> work, home -> theatre and theatre -> home (the reverse routes can be different because of one-way roads for example).

Adding more place to the map will result in a significant grow of combinations – according to permutations of n formula of combinatorics it can be calculated as:

A(k, n) = n! / (n - m)!

// where n - is a number of elements, 
//       k - is a number of elements in permutation (in our case k = 2)

Which gives us 12 routes for 4 places and 90 routes for 10 places (which is impressive). Note… this is without considering intermediate points between places i.e., to get from home to work you have to cross 4 streets, walk along the river and cross the bridge. If we imagine, some routes can have common intermediate points… well … as the result we will have a very large graph, with a lot of vertexes, where each vertex will represent either a place or a significant intermediate point. The class of problems, where we need to find all shortest paths between all pairs of vertexes in the graph, is called APSP (All Pairs Shortest Paths) and the base algorithm for solving these problems is Floyd-Warshall algorithm, which has O(n3) computational complexity.

And this is the algorithm we will implement today 🙂

Continue reading “Implementing Floyd-Warshall algorithm for solving all-pairs shortest paths problem in C#”