Multi-criteria evolutionary algorithm optimization for horticulture crop management

Climate variability requires adaptive production systems in agriculture often resulting in significant irreversible investments. Cultivar replacement programs in horticulture orchards that substitute older varieties for more heat- and drought-resilient varieties have enterprise values that are highly sensitive to the timing of such investments. Farm-level replacement programs are subject to multiple constraints around debt serviceability, operating costs, the replacement cycle and the rate of degradation of the existing orchard. The maximization of enterprise value subject to multiple constraints can be reduced to a multi-objective optimization problem. Over long horizons this optimization process generates a very-large solution space. Using a multi-objective evolutionary algorithm we examine uncertainties around climatic effects and the timing of investments for horticultural operations and derive the optimal times to adapt using cultivar replacement techniques. We find that the investment decision using traditional valuation methods is suboptimal and can result in poor decisions, potentially undermining adaptation efforts. We further show that opposing economic and climatic conditions can adversely impact enterprise value based on mistiming the investment decision. Application of the genetic algorithm solver is demonstrated using a vector-based geographic information system to a farm where individual portions of an orchard are subject to varying rates of production, degradation and age.