The cake eating problem is the simplest economic example of a finite dimensional dynamic programming environment. The problem can be described by the following:
- An agent lives through \(T\) periods and has preferences given the consumption of cake
- In each period \(t \in T\) the utility received by consuming \(c_t\) units of cake is represented by a function \(u(c_t) = ln(c_t + \alpha)\) where \(u(c_t)\) is monotone and convex
- Utility in future periods is discounted by a factor \(\beta \in (0,1)\) and cake in the first period is a non-zero endowment \(x_0 > 0\)
- The agent future consumption of cake \(c_{t+1}\) depends on what they previously chose to invest in the future \(x_t\)