WolframAlpha computational knowledge AI
max((0.5*w / (w+b)) + ((5-0.5*w)/(20-w-b))), 0 < w < 10, 0 <= b < 11