WolframAlpha computational knowledge AI
maximize 3*(a*(1-a)^2+b*(1-a-b)^2+c*(1-a-b-c)^2) subject to a+b+c<=1