In this study, genetic algorithm (GA) was utilized to optimize the phosphor distribution to decrease the junction temperature of white phosphor-converted light-emitting diodes (pc-LEDs). The key steps of the GA were introduced, including selection, crossover, and mutation. Both the junction temperature and the entransy dissipation of each evolution were calculated. It was found that with evolutions, the phosphor particles tend to build a "thermal bridge" between the chip and the convective boundary and spread along the convective boundary. The junction temperature decreases from similar to 157.5 degrees C to similar to 150 degrees C and the entransy dissipation decreases from 18 W K to 6 W K. The least entransy dissipation principle was demonstrated to be the rule that governs the optimization processes. (C) 2014 Elsevier Ltd. All rights reserved.