In Proof-of-Work, there are no rules on who will mine the next block - the selection is random and proportional to miners' computing power.
How is the next miner selected in various Proof-of-Stake block creation algorithms? Can the same miner / address / person be chosen to create multiple blocks in a row, or are there some mechanisms preventing this from happening?
Can the same miner / address / person be chosen to create multiple blocks in a row
Yes, a single entity can mine multiple blocks in a row with no ill effect.
are there some mechanisms preventing [multiple blocks in a row from one entity] from happening?
The sybil problem makes this fundamentally impossible; if there was a way to stop a person from pretending to be multiple entities at once, proof of stake and proof of work would not need to exist in the first place. In a naive attempt at this if you prevent a specific key from signing two blocks in a row, then the user would simple split their money into multiple keys.
How is the next miner selected in various Proof-of-Stake block creation algorithms?
In some implementations of proof of stake the selection is simply based on the previous block hash as a random oracle. ECDSA signatures can have a practically indefinite amount of permutations which allows a user in the network to simply sign repeatedly until the next "random" selection is also them, amplifying their income or ability to warp the network. This is commonly known as stake grinding.
In an attempt to fix this other implementations use either a block hash from a long time ago, or the hash of some period of blocks in the past. The intention of this is that an attacker would supposedly need to control all of the blocks in that period (or be extraordinarily lucky) in order to be able to re-sign and jump ahead with their new control.
There's other selection methods in existence but the enormous number of implementations make it extremely hard to research for this question, however they are likely variations on the same concept in some way. There's some more discussion of the general problems of proof of stake available for digestion in pos.pdf.
