From MGS to Global Transformations
- homepage
- abstract
- presentation, pdf, 6MB
- other material
- Fernandez A., Maignan L., Spicher A. (2019) Lindenmayer Systems and Global Transformations. In: UCNC 2019. LNCS 11493:65-78. doi:10.1007/978-3-030-19311-9_7
Subscribe to:
Post Comments (Atom)
Hi Antoine -- thanks for a very substantial presentation! One quick point of clarification. On slide 40, you have 0--3 cells. Does it stop there, or can you have n-D hypercells?
ReplyDeleteDear Susan, the mathematical object used here is abstract cellular complexes. Formally, it is a ranked partial order on the set of topological cells. In this context, the order relationship is interpreted as cells being incident or not, and the rank corresponds to the dimension of the cell. So to clearly answer, there is no theoretical limitation to the dimension.
DeleteSlide 33: A program is a set of rules. As the structure dynamically changes, there is the opportunitiy for new rules to come into effect / be applicaple (eg, in the case of embryo development -- as the embryo develops more, "higher level" processes start to occur. Are there any thoughts on how to make these *rules* dynamic/emergent? As you say on slide 5, "the structure of the state is needed to specify the evolution function" -- here I'm thinking not just specify as in select what's applicable here, but specify as in new state structure => new evolution function.
ReplyDeleteWhat you propose has been investigated in chemical computing (see Higher-Order Chemical Programming Style by J. -P. Banâtre, P. Fradet, Y. Radenac). In these works, rules are considered as first citizen data so that rules can be transformed as any other data. The programs is dynamic and rules can be generated by the program itself. At the moment, we do not have implemented higher-order features in MGS or in Global Transformations but it is an interesting suggestion of course.
DeleteWhat you propose has been investigated in chemical computing (see Higher-Order Chemical Programming Style by J. -P. Banâtre, P. Fradet, Y. Radenac). In these works, rules are considered as first citizen data so that rules can be transformed as any other data. The programs is dynamic and rules can be generated by the program itself. At the moment, we do not have implemented higher-order features in MGS or in Global Transformations but it is an interesting suggestion of course.
ReplyDelete