Abstract:

In order to solve the defects of poor training effects and low generalization ability of neural network in the prediction of concrete compressive strength， in this paper， the dolphin algorithm （DPO） and BP neural network are combined to construct a BP neural network model based on DPO（DPO-BP） and applied to optimize the concrete mixing ratio， and a set of feasible concrete mixing-ratio design schemes are proposed. The results show that compared with algorithms， genetic optimization BP neural network （GA-BP） and particle swarm optimization BP neural network （PSO-BP）， this optimization method has the advantages of fast convergence and good robustness. The DPO-BP algorithm has high solution accuracy and strong portability. It can also be used for slope stability judgment， machine fault diagnosis， air and water quality evaluation and many other fields. It has very significant engineering value.

Key words:

"> dolphin algorithm, BP neural network, concrete, compressive strength, mixing ratio