A Multiphase Semistatic Training Method for Swarm Confrontation Using Multiagent Deep Reinforcement Learning

Algorithm 1

Game rules.

	Input: Begin (Bool value)
	Output: Winner (if timeout or both teams are eliminated, return NULL)
(1)	Loop
(2)	Set
(3)	for and do
(4)	if, then
(5)	Break
(6)	else ifthen
(7)	Return
(8)	else ifthen
(9)	Return
(10)	end if
(11)	end for
(12)	Return