How you solve this depends on your available technology. Here's an old-school, low-tech version.
The weight of a single candy comes from a Normal distribution with Mean = 0.8538 and StdDev = 0.0518. You can first find the probability that a single random candy weighs less than 0.8532 as follows:
1) Convert to an equivalent z-score:
- z = (ActualValue - Mean)/StdDev = (.8532-.8538)/.0519 = -0.01156...
2) Use a z-score/Standard Normal table to find the probability of having a z-score less than this. If you are using a table, round the above z-score to -0.01; the corresponding value in the table is .4960.
Since the probability that a candy weighs less than the given amount is .4960, the probability that a candy weighs more than the given amount is 1 - .4960 = .5040.
Note that it's much easier, quicker, and more accurate to solve this with technology. If you can specify what you have available, you might get a more useful answer.