A refinement of the evolutionary theory is the doctrine of antagonistic pleiotropy, genes that have beneficial effects early and harmful effects late in the life history. For example, MMP13 is essential for very early development of collagen in mammals. Then the MMP pathway is quiescent. Later in life, the only beneficial effect known is to prevent gum disease. However the pathway is appropriated by the metastatic cascade, and plays a lead role in how cancer cells leave their initial site and penetrate the lymphatic system.

While the evolutionary theory of aging has come to reign as the accepted framework, workers in the field admit that even they don't have a ready understanding of its logic. Why don't all organisms evolve to reproduce as much as possible, as early as possible? As long as organisms continue to reproduce, why don't they continue to live, even indefinitely? Why do some organisms (such as primates) live significantly beyond reproductive peak? What drives organisms' widely varying lifespans such as a day for a mayfly, 200 years for a sea turtle, or a thousand years for a sequoia? How did senescence begin in our evolution on earth? What are the necessary and sufficient conditions for it to evolve? Are there one set of conditions, or many? Is sexual reproduction, which on earth coincided with the emergence of senescence, required, and if so, why?

Mathematics has illuminated vast spans of evolutionary biology, but only a small portion of the phenomenon of senescence. The population biology theory of age classes posits that a link between gene pathway activation and the age of the organism is a key to the evolution of senescence, which is certainly a cornerstone. The thesis of the current work is that computer simulations may offer a key to progress in understanding the logic of senescence from the standpoint of evolutionary theory.

We take an axiomatic approach to the conditions from which senescence may emerge. It is hypothesized that a minimal data structure required to capture senescence may look like this,

in which period is the clock of the system, followed by a list of genotypes, each consisting of the age class, the population (census), the probability that an organism in the age class will reproduce during the period, the likelihood of a mutation of the genotype into another genotype, the likelihood of mutation of the genotype's values for fecundity and longevity, the longevity or probability of living into the next period, and mortality, the probability that an extrinsic force will kill the organism during the period. Population biology does not separate the last two components (or axioms) and groups them together as survivorship.

The data structure must be coupled with a loop incorporating further minimal logic:

1. Increment the evolutionPeriod.
2. Increment all ageClasses.
3. Rotate left fecundity, mutationGenotype, mutationLM, and mortality to keep them with the same age class with which they were correlated before incrementing all age classes.
4. Multiply population in each age class x by longevity(x) to give the population that could live into the next generation (x+1) unless killed by mortality.
5. Multiply that product by mortality to give population that actually lives into the next generation (x+1).
6. Create a new age class if any population from any genotype's oldest age class survives another period.
7. Delete any maximum age classes in which census = 0.
8. Multiply the population in each age class x by fecundity(x) to give population in age class 0 at t+1.
9. Multiply the population in each genotype's age class 0 by the genotype mutation rate to give number of offspring to re-assign to other genotypes.
10. Multiply the population in each genotype's age class 0 by the fecundity-longevity mutation rate to vary the fecundity-longevity values for that genotype's offspring's age classes.

With this simple and intentionally minimal framework one can run many experiments in the evolution of senescence. Initial results of the experiments are presented. Limitations of the framework, likely extensions, implications for the connections between aging and disease, the relationship between previous theories of aging and the evolutionary theory, and implications for the engineering of longer lifespans are also presented.