Can parasites that copy twice as fast ever get rid of their hosts? Memory fills to max-creatures within about 50 ticks. After that every birth has to wait for the reaper to free space.
Try this Set mutation-rate to 0 and press Reset. Parasites never appear, so the soup stays all ancestors. Then set it back to 1 and watch the first diamonds spread.
About this model Where it ends up
Thomas Ray's Tierra (1991) ran self-replicating machine-code programs in a shared block of memory, with a scheduler handing out CPU time and a reaper removing old programs. Short parasites evolved that had lost their own copy loop and ran a neighbour's instead. This model keeps only that trade-off: copy faster, but depend on a host.
Each program is an agent with a genome size of 6 (an ancestor) or 3 (a parasite). Every tick an ancestor advances its copy by one step; a parasite advances only if an ancestor is within 2 cells. When the copy reaches the genome size, the program hatches a copy into a neighbouring cell, as long as there are fewer than max-creatures programs. With probability mutation-rate%, the copy's size is redrawn as 3 or 6. Programs die when older than reap-threshold ticks, or at half that age once memory is full.
There is no instruction set, no memory addressing and no evolution of genome content: a genome is just a size and a type. Squares are ancestors, diamonds are parasites, and colour shows age.
Idea Thomas S. Ray
MathPets MathPets · org
