How do I optimize my Dijkstra algorithm in Matlab to decrease the running time? - algorithm

Question

Given below is my Dijkstra algorithm. The B matrix is the cost map. The source node is R=[1,1] and there is no target node. I want to just compute the Dijkstra algorithm given the source node to all the nodes in matrix.
But my Dijkstra takes more than 20 minutes to run. I am not sure how to optimize my code. Can you please help me out here?
function hMap = djikstra(B,R)
%B = Cost Map
%R = Initial position of the robot
fprintf('In Dijkstra');
hMap = Inf(length(B)); % Assigning infinity
hMap(R(1),R(2)) = 0 ; % source node
open = [];
open = [R(1),R(2),0];
closed = [];
closed = [closed; open(1,1:2)]; %Put current source node in closed
x = R(1);
y = R(2);
while(size(open) ~= 0)
if(x-1 ~=0)
if(sum(ismember(closed(:,1:2),[x-1,y],'rows'))== 0)
cost = hMap(x,y)+ B(x-1,y);
fprintf('x-1');
%if(cost < hMap(x-1,y) )
hMap(x-1,y) = min(cost,hMap(x-1,y));
fprintf(' update x-1');
open = [open;x-1,y,hMap(x-1,y)];
%end
end
end
if(x+1 <= length(B))
if(sum(ismember(closed(:,1:2),[x+1,y],'rows'))== 0)
cost = hMap(x,y)+ B(x+1,y) ;
%if( cost < hMap(x+1,y))
hMap(x+1,y) = min(cost,hMap(x+1,y));
open = [open;x+1,y,hMap(x+1,y)];
%end
end
end
if(y-1 ~= 0)
if(sum(ismember(closed(:,1:2),[x,y-1],'rows'))== 0)
cost = hMap(x,y)+ B(x,y-1);
%if( cost < hMap(x,y-1))
hMap(x,y-1) = min(cost, hMap(x,y-1));
open = [open;x,y-1,hMap(x,y-1)] ;
%end
end
end
if(y+1 <= length(B))
if(sum(ismember(closed(:,1:2),[x,y+1],'rows'))== 0)
cost = hMap(x,y)+ B(x,y+1);
%if(cost< hMap(x,y+1))
hMap(x,y+1) = min(cost,hMap(x,y+1));
open = [open;x,y+1,hMap(x,y+1)];
%end
end
end
closed = [closed;x,y];
open = open(2:size(open,1),:); %Removing source element from the open list
open = unique(open,'rows');
open = sortrows(open,3); % Sorting w.r.t G Value
fprintf('------------------------------OPEN--------------------------')
open
if(size(open) ~= 0)
x = open(1,1)
y = open(1,2)
end
end
end