In this work, we propose the Networks of Evolutionary Processors (NEP) [2] as a computational model to solve problems related with biological phenomena. In our first approximation, we simulate biological processes related with cellular signaling and their implications in the metabolism, by using an architecture based on NEP (NEP architecture) and their specializations: Networks of Polarized Evolutionary Processors (NPEP) [1] and NEP Transducers (NEPT) [3]. In particular, we use this architecture to simulate the interplay between cellular processes related with the metabolism as the Krebs cycle and the malate-aspartate shuttle pathway (MAS) both being altered by signaling by calcium. NEP is complete and efficient from the computational point of view (i.e. is able to solve hard problems NP complete given linear time solutions). This model consists of several processors, each of one is placed in a node of a virtual graph. Each processor acts on local data in accordance with some predefined rules (evolutionary operations simulating point mutations of nucleotides over DNA sequences) and communicates the results using a filtering strategy. This strategy may require satisfy some conditions that are imposed by processors, when sending, receiving or both. The processors can communicate the resulting data with the rest of the processors connected with it. A processor node can be viewed as a cell that carries out only one specific evolutionary operation (substitution, elimination or insertion). NEPT uses these operations in order to generate recursively enumerable languages recognized by other NEP (without filtering strategy). On the other hand, NPEP uses these operations together with a valuation mapping (from strings to integers) to generate strings labeled with an electrical polarization. Each node has their own polarization, then the filtering strategy consists in let pass those strings with their same polarization.
展开▼
