1. H. Lifson's teams model of the molecube in simulation showing distinct replication stratergies in physical simulation. This is an example of how a non-uniform cellular automata can be used to explore the intrinsic fitness of rule sets. It is interesting that their equation for self-replication does not include the cooperative cross-catalytic self-replication of multiple rule sets that are not spatially connected. This may miss many distributed kinds of self-replication. The concept of distributed self-replication should be properly developed using my template model if possible.
2. H. Williams' model of flask based surface evolution in a population of replicators utilizing resources and under micromutation, with energy conservation of sorts. The aim is to see to what extent 'ecosystems globaly self-maintain themselves'. I am not sure how this differs from Law's community assembly models? I am also not sure how the hypothesis can be tested operationally. The model of metabolism and membrane co-evolution should have similar properties to this in many respects, except the physical properties of the chemicals will allow a greater capacity for surface organizations to arise.
3. H. Bersini et al's model of artificial chemistry. Rather complex graph grammer for a molecule, too complex for my purposes perhaps, but potentially capable of considerable variety in chemical topology. i have to produce a simpler version of this, and explore the capacity of a node level structure description to generate interesting graph structures in the artificial chemistry. Recently I found Roy Johnston doing work on theoretical chemistry at Birmingham University. Hopefully we will share similar interests in the metabolism model. The details of this work can be found here.
4. Hofstadter's claim for the failure of Moore's Law w.r.t. computing power. I wonder to what extent energy constraints have played a crucial role in the development of technology and how this can be tested. To what extent does search depend on energy? To what extent is random search cheaper energetically than developing a stratergy for non-random search, i.e. what is the energetic cost of evolution and learning? Perhaps it would be useful to write a paper considering the means by which an entity driven by one prime mover is able to discover another prime mover, ranging from bacteria to humans. This is a historical and biological piece of work that would require close scholerly examination of the processes involved in the discovery of new prime movers, and the costs of this search! Hofstadter was unimpressive here, and as I had suspected, he does not take physical systems on the chin, but ducks and dives from them. The contribution that he can make to issues of pre-bioitic chemistry thus seems limited, even through in GEB he claims that the concept of self-referential activity is crucial to life. He will not quite venture into the unknown in the way that a person of his stature should be willing to do. I make a start by considering the importance of energy in open-ended evolution here. I hope to be able to tell a general story of how energy is intimately linked to evolution.
5. L. Rocha's work on RNA editing may relate to my idea about ribozyme re-writing grammers. I should hurry up and work on this a bit more quickly to see what sorts of processing a system capable of ribozyme type cleavage and ligation reactions is capable of. The ribozymes would be evolved to carry out certain dynamic functions using cleavage and ligation reactions. Signals from the cell surface would result in the production of certain ribozymes, and these ribozymes would be able to act on other ribozymes, either ligating or cleaving them in a sequence specific manner. As a control the same functions would be evolved with a CTRNN! We would compere the solutions used to solve this problem in both cases. This would fit in nicely with the cell signaling network remit.
6. H. Susuki et al are working on models of the co-evolution of metabolism and membrane on surfaces. This is precisely what I want to do, but with a more meaningful and realistic model of chemistry that suffers from the problem of side-reactions. The particular question I would like to answer is whether the existence of surfaces with chemicals posessing physical properties can aid the self-organization of increasingly complex chemical organizations! One simple way to approach this problem is to consider first a random chemistry on a surface, i.e. where particles undergo rare stochastic events, and these react with the chemicals currently in existence in proportion to the free energy differences between the chemicals in existence. Obviously we do not want to allow all possible reactions to take place, and so we impose constraints based on sharing of complementary active sites. Further work on this model can be found here.
7. Simon McGregor has had the excellent idea of finding applications of Alife and AI to 3rd world problems. The field seems ripe for consideration. I would love to hear from people who are interested in this, and would like to help take it forward.
8. Randel Beer type research is suited to help me understand CSNs. CTRNNs can be used to model CSNs, and the tasks they have to undertake. Different types of transfer function etc.. can be investigated to see how easily they are able to solve CSN type problems. I need to look up cases of learning in bacteria, or learning at the cell level in multi-cellular organisms. Then implement these tasks in a CTRNN.