A multi-objective optimization platform for artificial lighting system in commercial greenhouses

Abstract Limited natural daylight in Nordic Countries means artificial lighting is a critical factor industrial plant production. The electricity cost of lights accounts for large percentage the overall optimal use production can be formulated as multi-objective problem (MOP) to achieve growth while minimizing cost. In previous work, solving this MOP, Genetic Algorithm (GA) was used create Pareto Frontier (PF), which contains solutions representing trade-off using against objectives. PF updated immediately once non-dominated child-solution found by comparing dominance with PF. Besides, addition PF, initial random population also reused parent source evolution process. When genetic process terminated, priority-based selection mechanism select final solution from paper, an alternative strategy proposed and compared GA strategy. By strategy, all child-solutions are only their parents during process, collected into candidate list. then at end each generation on solutions. parent-solution addition, posterior normalization implemented evaluation, social welfare metrics (SWs) applied avoid explicit ranking experimental results show that outperforms dramatically avoiding local minima.