Roadmap of Investigations

To reconstruct the concept of a LCS again we will following the following general 'roadmap':

1. Prepare a statistical model for the wood1 as well as wood1-similar environments (cf. Wilson 1994 [347]).
2. Define a first set of simple tasks to be solved (this represents environment $E_{0}$).
3. Prepare a purely random agent $ANIMAT^{O}$.
4. Show the performance of agent $ANIMAT^{O}$ for these tasks.
5. Construct an enhanced agent $ANIMAT^{1}$ with a complete reactive behavior for every possible perception.
6. Show the performance of this $ANIMAT^{1}$ compared to $ANIMAT^{O}$.
7. Construct an enhanced agent $zcs$ from Wilson 1994 [347] with GA-like learning based on an external fitness function.
8. Show the performance of this $zcs$ compared to $ANIMAT^{O}$ and $ANIMAT^{1}$.
9. Construct an enhanced agent $ANIMAT^{2}$ with a 2-level memory and a built-in fitness function. The GA-like learning uses the built-in fitness function as well as the memory, which can show actions and their results from the past organized in different paths.
10. Show the performance of this $ANIMAT^{2}$ compared to all the other preceding agent types.
11. Construct an enhanced agent $ANIMAT^{3}$ with a 4-level memory and a built-in fitness function. The GA-like learning uses the built-in fitness function as well as the memory, which can show actions and their results from the past organized in different paths on the levels 1-2. On level 3 the memory represents non-language objects $O_{\overline{L}}$ as well as language objects $O_{L}$. On level 4 the memory represents sign relations mapping language objects onto non-language objects and vice versa $\rho : O_{L}^{n} \longmapsto 2^{O_{\overline{L}}^{n}}$.
12. Show the performance of this $ANIMAT^{3}$ compared to all the other preceding agent types.