Abstract. Nitrogen-fixing bacteria (rhizobia) have symbiotic relationships with legumes: they inhabit legume root nodules and convert atmospheric nitrogen to a plant available form in exchange for photosynthates. Generally, this symbiotic process called biological nitrogen fixation is mutually beneficial to both plants and bacteria. Using this mechanism symbionts acquire alternative sources of hard-to-reach individual growth resources (carbon for rhizobia and nitrogen for plants). However, not all rhizobia provide fixed nitrogen to the host plant honestly: some of them can behave as a kind of cheaters. Unlimited cheating rhizobia strains propagation may potentially disrupt the symbiotic relationships. This raises the question of plant–rhizobia mutualism evolutionary stability. This paper presents the results of the legume–rhizobia interactions investigation implemented as AnyLogic agent-based models. Three modifications of interaction model (“one plant – one strain of rhizobia”, “one plant – several strains of rhizobia”, “one plant with root nitrogen uptake – several strains of rhizobia”) in the form of evolutionary games in two populations (rhizobia and plants) are considered by the authors. Simulated natural selection is driven by populations heterogeneity: each agent has its own cooperation parameter which determines its strategy in evolutionary game. In the set of numerical experiments the following results were obtained. Simulated populations tend to become homogeneous with cooperation parameter value close to the theoretically optimal. Such degenerated structure of populations is evolutionarily stable and maximizes the total growth of the entire symbiotic system. Thus, the logic of symbionts co-development simulation itself prevents the emergence of parasitic strategies and automatically provides rational and mutually beneficial partnership sustainability. This remains true in the early stages of ontogenesis or under the assumption that life cycle duration is unlimited.

 

Key words: biological nitrogen fixation, agent simulation, evolutionary game theory, evolutionarily stable strategy.