Updated Divide and Conquer framework

The Divide and Conquer framework (for decomposed discovery and decomposed replay) now also supports a move-on-model activity.

A transition that is mapped to this activity will always result in a visible move-on-model in the replayer, like a transition that is mapped to the invisible activity will always result in an invisible move-on-model. The difference between both is the costs between both moves: An invisible move-on-model has typically no costs associated with it, while a visible move-on-model has. By mapping a transition to the move-on-model activity, one now makes sure that it it needs to be executed, it has costs associated with it, without having to assign an existing activity to it.

The (maximal) decomposition of the net handles the move-on-model activity by considering it to be an activity carrying the transition label. If a transition with unique label is mapped onto the move-on-model activity, the decomposition may split subnets on this transition. Subnets containing transitions with non-unique labels mapped onto the move-on-model activity will be merged into a single subnet, like happens with other transitions that are mapped onto the same activity.