Maximum Clique
People in the social network are identified by unique IDs, consecutive integers from 1 to N. Who follows who is captured in a cell array called sn: the ii th element of sn is a vector that contains a list of IDs the person with ID ii follows. You may assume that these lists are ordered in ascending order by ID. Note that the follows relationship is not necessarily symmetrical: if person A follows person B, person B may or may not follow person A. :
function clique = max_clique(graph, clique)
if nargin < 2
clique = [];
end
max_clq = clique;
if isempty(clique)
for ii= 1:length(graph)
clq = max_clique(graph,ii);
if length(clq) > length(max_clq)
max_clq = clq;
end
end
else
for node=1:length(graph)
if isempty(find(node==clique))
if check_clique(clique,node,graph)
clq = max_clique(graph, [clique node]);
if length(clq) > length(max_clq)
max_clique == clq
end
end
end
end
end
clique = max_clq;
end
function ok = check_clique(clq,node,graph)
ok = false;
for ii=1:length(clq)
if isempty(find(clq(ii) == graph{node})) || isempty (find(node == graph{clq(ii)}))
return;
end
end
ok = true;
end
Unfortunately, it is too slow and the grader will time out. Your task is to modify the code to speed it up. Remember, the question to ask: am I doing any unncessary work? Call the modified function max_clique. (Hint: when we try to expand the current clique, do we really need to consider all the nodes?)
Please solve this problem with entire new code that is fast.

9 个评论

Here is the mat file with the example social network: sn.mat
Since this is a homework assignment, show us the code you've written to try to solve the problem and ask a specific question about where you're having difficulty and we may be able to provide some guidance.
If you aren't sure where to start because you're not familiar with how to write MATLAB code, I suggest you start with the MATLAB Onramp tutorial (https://www.mathworks.com/support/learn-with-matlab-tutorials.html) to quickly learn the essentials of MATLAB.
If you aren't sure where to start because you're not familiar with the mathematics you'll need to solve the problem, I recommend asking your professor and/or teaching assistant for help.
I am not able to understand the problem Can show me the road map to solve it
The auto grader is timimg out with my solution Can you give a more efficient algorithm
Finding a way to improve the performance of the code is literally your assignment. "Unfortunately, it is too slow and the grader will time out. Your task is to modify the code to speed it up." We will not write the code for you.
If you're not sure what to do in terms of algorithm, contact your professor and/or teaching assistant.
I know its my task my task but i am not able to contact the teacher
can you atleast study the code and give me tips
This line in the posted code is not useful:
max_clique == clq
This is a comparison, but it does not assign anything. In the screenshot of your code this line is:
max_clique = clq;
Which of the two versions is wanted? Please do not let the readers guess.
Remember that:
isempty(find(a == b))
can be simplified by
~any(a == b)
You get corresponding hints in the editor already. Do not ignore them.
Jan
Jan 2021-4-3
编辑:Jan 2021-6-16
@Parth Tushar Deodhar: You have set a flag:
"please delete this thread as it is an home work assignment and might lead to some one copying it."
With using this forum you agree to the Terms of Use. This implies that the posted contents is made available for the community. The nature of this forum is the sharing of problems and solutions. To support this, many voluntary Matlab users spend their time. Removing a question after an answer has been given, would not respect their work.
Please think twice before you post homework questions in the internet. Remember that even if the thread is deleted in the forum, you will still find a copy e.g. in Googles cache. Therefore I remove the flag without deleting the thread.
If you reach a code send it here to me ???

请先登录,再进行评论。

 采纳的回答

Jan
Jan 2021-3-15
编辑:Jan 2021-3-15
With replacing
if isempty(find(node==clique))
...
if isempty(find(clq(ii) == graph{node})) || isempty (find(node == graph{clq(ii)}))
by
if ~any(node==clique)
...
if ~any(clq(ii) == graph{node}) || ~any(node == graph{clq(ii)})
the processing time is reduced from 350 seconds to 193 seconds on my Matlab R2018b.
As next step inline the frequently called function check_clique in the main function:
function clique = max_clique(graph, clique)
if nargin < 2
clique = [];
end
max_clq = clique;
if isempty(clique)
for ii= 1:length(graph)
clq = max_clique(graph, ii);
if length(clq) > length(max_clq)
max_clq = clq;
end
end
else
for node = 1:length(graph)
if ~any(node == clique)
ok = true; % Inlined check_clique:
for ii = 1:length(clique)
if ~any(clique(ii) == graph{node}) || ...
~any(node == graph{clique(ii)})
ok = false;
break;
end
end
if ok % check_clique(clique,node,graph)
clq = max_clique(graph, [clique, node]);
if length(clq) > length(max_clq)
max_clq = clq;
end
end
end
end
end
clique = max_clq;
end
This needs 72 seconds on my computer. 5 times faster with just tiny modifications.

11 个评论

The AutoGrader is still timimg out and asking for a better and more efficient alogorithm with a hint that do we need to consider all nodes when we try to expand the current clique. Please Help
Are you aware that you didn't mention yet what the code should calculate?
Hello Jan, I have tried the code however it still takes long computation time.
Yes, of course it is not much faster. I've suggested some obvious improvements only, but did not modify the algorithm in any way. The OP forgot to mention, what the code should compute. I stay away from guessing this very important detail.
Now you run the code again. Why? Did you get the same homework? If so, how can we help you?
Hello @Parth Tushar Deodhar, @Fatema Saeed, I hope you have successfully solved the problem.
I have been struggling with this assignment for quite a while. I think what the code should compute is a maximal social network clique where everyone follows everyone else. The original version is inefficient because it is checking a bunch of sets of nodes that cannot possibly be part of a clique. So we need a smarter way to pick which nodes to check.
Therefore, based the above tiny improvement, I also tried:
  1. disregard nodes with length < length(clique)
  2. disregard nodes with the first element > clique(1)
  3. disregard nodes with the last element < clique(end)
  4. disregard nodes that are not in the follow list of first elemnt in the current clique
So my version of modification based on what @Jan wrote is changing line 15 to:
if ~any(node == clique) && length(graph{node}) >= length(clique) && graph{node}(1) <= clique(1) && graph{node}(end) >= clique(end) && any(node == graph{clique(1)})
Unfortunately, this runs for ~60 seconds on my computer, which is still not fast enough to pass the grader...
I also come up with the order of operands of && operator. So I change the above line to:
if any(node == graph{clique(1)}) && length(graph{node}) >= length(clique) && graph{node}(1) <= clique(1) && graph{node}(end) >= clique(end) && ~any(node == clique)
Although this gets a bit faster and runs for ~40 seconds on my computer, unfortunately this is still not fast enough to pass the grader...
I REALLY HAVE NO FURTHER IDEAS.
I HOPE SOMEONE COULD PROVIDE SOME ADVICES.
ANY SUGGESTIONS WILL BE APPRECIATED.
I want a code that can help me????
function clique = max_clique(graph, clique)
if nargin < 2
clique = [];
end
max_clq = clique;
if isempty(clique)
for ii= 1:length(graph)
clq = max_clique(graph,ii);
if length(clq) > length(max_clq)
max_clq = clq;
end
end
else
for node=1:length(graph)
if ~any(node==clique)
if check_clique(clique,node,graph)
clq = max_clique(graph, [clique node]);
if length(clq) > length(max_clq)
clique = max_clq;
end
end
end
end
end
clique = max_clq;
end
function ok = check_clique(clq,node,graph)
ok = false;
for ii=1:length(clq)
if ~any(clq(ii) == graph{node}) || ~any(node == graph{clq(ii)})
return;
end
end
ok = true;
end
i make this code but not run because of the time run out
can anyone help me to solve this problem?
I suspect you could write the code more efficiently using ismember() in check_clique
anyone reach the answer of this code ??

请先登录,再进行评论。

更多回答(5 个)

function clique = max_clique(g,clique)
if nargin < 2
clique = [];
end
max_clq = clique;
if isempty(clique)
for ii = 1:length(g)
clq = max_clique(g,ii);
if length(clq) > length(max_clq)
max_clq = clq;
end
end
else
candidates = g{clique(1)};
candidates = candidates(g{clique(1)} > max(clique));
for ii = 1:length(candidates)
if check_clq(clique,candidates(ii),g)
clq = max_clique(g,[clique candidates(ii)]);
if length(clq) > length(max_clq)
max_clq = clq;
end
end
end
end
clique = max_clq;
end
function ok = check_clq(clq,id,g)
ok = false;
if ~isempty(find(id == clq))
return;
end
for ii = 1:length(clq)
if isempty(find(clq(ii) == g{id})) || isempty(find(id == g{clq(ii)}))
return;
end
end
ok = true;
end
Mehrail Nabil
Mehrail Nabil 2021-8-17

0 个投票

Anyone can send me in comment the answer of it???? The code please

2 个评论

@Mehrail Nabil Were you able to solve the problem? I would appreciate if you help me with the code. Please.
function clique = max_clique(g,clique)
if nargin < 2
clique = [];
end
max_clq = clique;
if isempty(clique)
for ii = 1:length(g)
clq = max_clique(g,ii);
if length(clq) > length(max_clq)
max_clq = clq;
end
end
else
candidates = g{clique(1)};
candidates = candidates(g{clique(1)} > max(clique));
for ii = 1:length(candidates)
if check_clq(clique,candidates(ii),g)
clq = max_clique(g,[clique candidates(ii)]);
if length(clq) > length(max_clq)
max_clq = clq;
end
end
end
end
clique = max_clq;
end
function ok = check_clq(clq,id,g)
ok = false;
if ~isempty(find(id == clq))
return;
end
for ii = 1:length(clq)
if isempty(find(clq(ii) == g{id})) || isempty(find(id == g{clq(ii)}))
return;
end
end
ok = true;
end

请先登录,再进行评论。

MR MB
MR MB 2021-9-4
编辑:MR MB 2021-9-4
%if true
function clique = max_clique(graph,clique)
if nargin < 2 % originaly we call the function with just the graph
clique = []; % hence, the clique is initialized to an empty vector
end
max_clq = clique; % max_clq will store the current largest clique
if isempty(clique) % when we first call the function
ii = 1:length(graph);
s = max_clique(graph,ii); %out of the loop
for ii = 1:length(graph) % we need to test potential cliques starting from every possible node
clq = s;
if length(clq) > length(max_clq) % if the new one is larger than the current
max_clq = clq; % we store the new one
end
end
else
for node = 1:length(graph) % we are in a recursive call now: we test every node as a new member
if isempty(find(node == clique)) % unless it is already in the clique
if check_clique(clique,node,graph) % if adding this node is still a clique
clq = max_clique(graph,[clique node]); % we call ourself with the new expanded clique
if length(clq) > length(max_clq) % if what we get is larger the curent max
max_clq = clq; % we store the new one
end
end
end
end
end
clique = max_clq; % return the largest one
end
%if true
function ok = check_clique(clq,node,graph) % adding node to clq still a clique?
ok = false;
for ii = 1:length(clq) % for every current member
if isempty(find(clq(ii) == graph{node})) || ... % the member must be on the follows list of the new guy
isempty(find(node == graph{clq(ii)})) % the new guy must be on the follows list of the member
return;
end
end
ok = true;
end
end
It is so so so fast but with wrong answer Can any one help me to find the mistake?!?

5 个评论

anyone here can help me? I have problem and I get this Error
Undefined function 'max_clique' for input arguments of type 'cell'.
this is my code:
functuin clique = max_clique(graph,clique)
if nargin<2
clique=[];
end
max_clq=clique;
if isempty(clique)
for ii=1:length(graph)
clq=max_clique(graph,ii);
if length(clq)>length(max_clq)
max_clq=clq;
end
end
else
for node=1:length(graph)
if ~any(node==clique)
ok=true;
for ii=1:length(clique)
if ~any(clq(ii)==graph{node})||...
~any(node==graph{clq(ii)})
ok=false;
break;
end
end
if ok
clq=max_clique(graph,[clique,node]);
if length(clq)>length(max_clq)
max_clq=clq;
end
end
end
end
end
clique=max_clq;
Have you seen this: "functuin"?
Thanks Jan, but now I get this error
Undefined function 'max_clique' for input arguments of type 'cell'.
Thanks alot friend my problem solved

请先登录,再进行评论。

function clique = max_clique(g,clique)
if nargin < 2
clique = [];
end
max_clq = clique;
if isempty(clique)
for ii = 1:length(g)
clq = max_clique(g,ii);
if length(clq) > length(max_clq)
max_clq = clq;
end
end
else
candidates = g{clique(1)}; % it is enough to check nodes that the first member of the clique follows
candidates = candidates(g{clique(1)} > max(clique)); % since nodes are ordered, a potential new member must have a greater ID than current members
for ii = 1:length(candidates)
if check_clq(clique,candidates(ii),g)
clq = max_clique(g,[clique candidates(ii)]);
if length(clq) > length(max_clq)
max_clq = clq;
end
end
end
end
clique = max_clq;
end
function ok = check_clq(clq,id,g)
ok = false;
if ~isempty(find(id == clq))
return;
end
for ii = 1:length(clq)
if isempty(find(clq(ii) == g{id})) || isempty(find(id == g{clq(ii)}))
return;
end
end
ok = true;
end
function clique = max_clique(g,clique)
if nargin < 2
clique = [];
end
max_clq = clique;
if isempty(clique)
for ii = 1:length(g)
clq = max_clique(g,ii);
if length(clq) > length(max_clq)
max_clq = clq;
end
end
else
candidates = g{clique(1)}; % it is enough to check nodes that the first member of the clique follows
candidates = candidates(g{clique(1)} > max(clique)); % since nodes are ordered, a potential new member must have a greater ID than current members
for ii = 1:length(candidates)
if check_clq(clique,candidates(ii),g)
clq = max_clique(g,[clique candidates(ii)]);
if length(clq) > length(max_clq)
max_clq = clq;
end
end
end
end
clique = max_clq;
end
function ok = check_clq(clq,id,g)
ok = false;
if ~isempty(find(id == clq))
return;
end
for ii = 1:length(clq)
if isempty(find(clq(ii) == g{id})) || isempty(find(id == g{clq(ii)}))
return;
end
end
ok = true;
end

1 个评论

This is the same as what @Rajith posted, except for using slightly different number of spaces at the beginning of lines. There is no difference other than the spacing.

请先登录,再进行评论。

类别

帮助中心File Exchange 中查找有关 Logical 的更多信息

产品

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by