Today I found out about amoeba based computing and optic computing in a popular math book In Pursuit of the Traveling Salesman: Mathematics at the Limits of Computation by William J. Cook. Searching more online In Amoeba-Based Emergent Computing:Combinatorial Optimization and Autonomous Meta-Problem Solving the authors conclude:
When applied to some techniques in automated reasoning and machine learning [34],meta-problem solving may be a useful concept to create autonomous agents that are capable of coping with unexpected situations. Suppose that our computing system is developed to be embedded in a control system of an autonomous agent such as a mobile robot. The standard problem-solving process to find a solution to a given problem would correspond to the agent’s computing process to select an appropriate behavior in an expected situation by following instructions of a given program prescribed in advance by a programmer. However, the program prescribed only for expected situations would always be insufficient for the agent to act in an actual environment where unexpected situations arise very often. The agent may unexpectedly be stuck in a stalemated situation in which the given program is helpless for finding a way out of the stalemate. In such a case, if a meta-problem solver was incorporated in the agent’s control system, the agent evolves the program and may succeed in finding an emergent solution to overcome the stalemate owing to the meta-problem solver’s capability of the spontaneous destabilization. This is a desirable scenario that we picture in our minds as a potential application of meta-problem solving.
But the resources (there is not even a Wikipedia page) on this is sparse.
- Is there a formal treatment of amoeba-based computing (abc) that attempts to prove some metamathematical theorems other than TSP?
- How is cellular automata (CA) used to model amoeba?
Some Useful Links: