consider the following snippet in cpp. This is taken from dynamic programming tutorial .
It is mainly for space optimized knapsack problem.
for(int i=1;i<=a;i++) { int t = a%2; for(int j=0;j<1100000;j++) dp[t][j] = inf; for(int j=0;j<1100000;j++){ dp[t][j] = min(dp[t][j],dp[!t][j-v[i-1]]+w[i-1]);//!t part is for skipping the current } }
This snippet is taken from this tutorial. I want to convert this technique into java. But java does not support this type of integer manipulation. Please can anyone explain me how it works and appropriate conversion to java ? thanks.
Advertisement
Answer
Just replace !t
with t ^ 1
or 1 - t
(whatever you find more readable; the effect is the same). That’s all you need to do here.
Explanation
Java supports all the integer manipulation on display in this snippet:
int t = a % 2;
<– this is valid java and means the same thing in java as it does in C: divide a by 2, and put the remainder in t, preserving the sign of a
. Thus: t
is now 0 if a
was even, and 1
if a was positive and odd, and -1
if a was negative and odd. It looks like a
is supposed to be positive in this scenario, meaning that t
can only be 0
or 1
here.
dp[t][j]
<– valid java. Declare dp
as for example int[][] dp = new int[100][100];
.
min(someExpression, someOtherExpression)
<– valid java; add import static java.lang.Math.min;
or replace min with Math.min
to make it work.
dp[!t]
<– the !t
isn’t valid java; but, in C, running !t
where t
is either 0 or 1 is just flipping things: If t
is 1, !t
is 0, and vice versa. And that you can trivially do in java: Use t ^ 1
or 1 - t
or even t == 0 ? 1 : 0
– all have the exact same behaviour and are valid java.