This is an algorithmic issue up to a point. Local optimization is often 'easier' than global optimization but does not necessarily find a global optimum. On th

This is an algorithmic issue up to a point. Local optimization is often 'easier' than global optimization but does not necessarily find a global optimum. On the other hand, some systems do just fine navigating a complex problem space by just solving large number of local optimization problems. See for example slime molds, which are colonies without centralized nervous system, yet manage to move and forage for sustenance successfully as long as the environment is solvable. It's more efficient if everyone get's their own breakfast but the central bank should take care of the interest rate affecting the money that's used to buy the breakfast. Etc