In layman's terms, not in programming terms, the algorithms are all used for sorting data. When you are in Excel and want to sort the values in a column by going to Data->Sort, it's using one of these algorithms or something similar.
The difference between them is the approach they take to sorting. One just compares adjacent items, resorting them one at a time until the whole list is sorted, one splits the list into two parts, splits the parts into two each, all the way until you can't sort any more, then sorts the split up pieces and stiches them back together. There are other strategies in the other algorithms.
Obviously they're all best used for sorting data, but you probably want a more specific answer. That would require getting into some technical detail that would go beyond layman's terms. I can say that the details of some algorithms make them better for sorting some types of data sets - each algorithm has a "best case", "worst case", and "expected case". If yourset of data is expected to be similar to an algorithm's "best case", then you'll probably get good performance by using that algorithm.
When you see some people refer to the complexity of an algorithm with an expression like "O(n log n)", that's computer science notation. It means that on average, we expect this algorithm to take n * log n units of time or steps to complete. n in the case of sorting algorithms is the number of pieces of data to be sorted. log is a base 10 logarithm (pull out your old algebra book!). Basically, the big-O notation allows us to talk about the performance of an algorithm independent of the hardware or sotware that it's implemented on.
Good luck.
Jay