Emergency scheduling optimization based on improved artificial bee colony algorithm

When large scale disaster occurs suddenly, allocating scheme of the first batch of "life" emergency supplies directly determines rescue effect. To make lack loss evaluation of the supplies accordant with actual cases, this paper proposed nonlinear loss evaluation function, based on which constructed multi to multi constrained scheduling model with multi objectives. To improve the performance of artificial bee colony algorithm, this paper proposed following schemes: food sources initialization based on converse food source.; reverse learning and comprehensive learning added into bee searching procedure. Experiments results on random data of three scales show that non dominated front solutions set solved by the improved algorithm is more diverse, wide spread and uniform.