Computational Budgets on multi-action turns

[hopshackle]

In many games once can make a number of distinct actions as part of a 'Turn'. Pandemic is an example, where one has three actions per turn - and some of those actions (e.g. to play a card) then open up additional sub-actions. Dominion similarly can have one turn involving the play of multiple cards.

These sorts of multi-action games have led to Online Evolutionary Programming (OEP), Evolutionary MCTS (eMCTS), Bridge-Burning MCTS and so forth. The central underlying point of these is that we have a single computational budget that has to be spent over a number (fixed, or possible undefined at the start) of distinct actions. The collectivity of all these actions is then a single 'Turn'.

For future enhancement in TAG would be some way to support this idea in a game.