Abstract We present a cellular automaton that simulates the interaction between a host tree and multiple potential mycorrhizal symbionts and generates testable hypotheses of how processes at the scale of individual root tips may explain mycorrhizal community composition. Existing theoretical biological market models imply that a single host is able to interact with and select from multiple symbionts to organize an optimal symbiont community. When evaluating the tree–symbiont interaction, two scales must be considered simultaneously: the scale of the entire host plant at which carbon utilization and nutrient demands operate, and the scale of the individual root tip, at which colonization and carbon-nutrient trade occurs. Three strategies that may be employed by the host tree for optimizing carbon use and nutrient acquisition through mycorrhizal symbiont communities are simulated: (1) carbon pool adjustment, in which the plant controls only the total amount of carbon to be distributed uniformly throughout the root system, (2) symbiont selection, wherein the plant opts either for or against the interaction at each fine root tip, and (3) selective carbon allocation, wherein the plant adjusts the amount of carbon allocated to each root tip based on the cost of nutrients. Strategies were tested over various nutrient availabilities (the amount of inorganically and organically bound nutrients). Success was defined on the basis of minimizing carbon expended for nutrient acquisition because this would allow more carbon to be utilized for growth and reproduction. In all cases, the symbiont selection and selective carbon allocation strategies were able to meet the nutritional requirements of the plant, but did not necessarily optimize carbon use. The carbon pool adjustment strategy is the only strategy that does not operate at the individual root tip scale, and the strategy was not successful when inorganic nutrients were scarce since there is no mechanism to exclude suboptimal symbionts. The combination of the symbiont selection strategy and the carbon pool adjustment resulted in optimal carbon use and nutrient acquisition under all environmental conditions but result in monospecific symbiont assemblages. On the other hand, the selective carbon allocation strategy is the only strategy that maintained successful, multi-symbiont communities. The simulations presented here thus imply clear hypotheses about the effect of nutrient availability on symbiont selection and mycorrhizal community richness and composition.