Fig. 6
From: Learning versus optimal intervention in random Boolean networks
The scalability of the optimal rule algorithm. Each point represents a particular network instance with given parameter values of N and K. Uncovering the optimal rules for each network (in terms of minimal control cost) takes a particular amount of time (as shown on the x-axis) and produces a certain number of control rules (as shown on the y-axis). Network instances that have a value of N equal to 11 or less take relatively little time to run (less than 200 s). Network instances with a value of greater than 11 may take significantly longer to run or may fail to complete due to RAM limitations