1. Counting sort and radix sort can sort in linear time O(n) by exploiting properties of the input rather than just comparisons. Counting sort assumes integers as input and radix sort assumes digitized numbers.
2. Bucket sort also runs in linear time if inputs are uniformly distributed between 0 and 1. It divides the range into buckets and distributes inputs into the corresponding buckets which are then sorted.
3. The comparison-based lower bound of Ω(nlogn) does not apply to these algorithms because they do not rely solely on comparisons. Counting sort counts occurrences rather than comparing, and radix/bucket sort distribute into buckets based on digits/positions rather than comparisons.