In this study, we address the integrated scheduling problem involving quay cranes and IGVs in automated terminals. We construct a mixed-integer planning model with the aim of minimizing the total energy consumption during quay crane and IGV operations, focusing on the loading-operation mode. The model considers the impact of the actual stowage of container ships on the loading order. We propose a dimension-by-dimension mutation sparrow search algorithm to optimize the model’s solution quality. Building upon the standard sparrow search algorithm, we incorporate cat mapping to enhance the diversity of the initial sparrow population. To improve global search in the early stage and local search in the later stage of the algorithm, we introduce an adaptive t-distribution mutation strategy. Finally, a total of 12 instances with container counts containing 30, 100, and 250 were designed for experiments to validate the effectiveness of the model and algorithm. The experiments demonstrate that, by appropriately increasing the number of quay cranes, configuring more than two or three IGVs can achieve optimal energy consumption for overall operations.

To solve the multi-task assigement and routing optimizating problem of AGVs, this paper proposed a hybrid algorithm. Based on analysis of a practice case, it divided the working process of AGVs into two stages: loading and unloading. In the first stage, the shorts paths between loading site and unloading site were obtained via Dijkstra method. Then in next stage, a pre-processing method was applied to simplify the path network. Then the unloading travel distance was optimized by Hungarian algorithm. Through the two-section processing method, the optimal of routing design could be obtained. At last, the result of a practice case showed that the method could be used to solve the routing optimization of AGVs with multi-task assignment problems.

This paper proposed a navel encoding strategy with multiple segments of swarm intelligence optimization algorithms to solve the set-covering location-allocation problem with collaborative inventory. It divided an individual expression into several segments and then decoded the location and allocation information. By this encoding strategy, the discrete and continuous problems can be analysed simultaneously. It could improve the search efficiency of the swarm intelligence optimization algorithms to solve the problem. The experiment results showed that the strategy was viable and could be applied for the selected algorithms to optimize the set-covering location-allocation problem with collaborative inventory.

In order to apply swarm intelligence optimization algorithms to solve cold chain vehicle routing problem efficacy and conveniently, a simplified decoding method was proposed. Based on the common encoding form in continuous space for swarm intelligence optimization algorithms, it divided the decoding process into three sections: decoding the service sequence of customers, assigning customer and determining the route of each vehicle, calculating circularly and outputting the whole plan. It did not need to design new encoding form and executed efficacy. Experiments with two different scale problem were introduced to test the performance of the decoding method applied in 9 algorithms. The results showed that the method can be used in swarm intelligence algorithms to solve the cold chain logistics vehicle routing problem. In addition, it also can be seen that different algorithms showed different performances due to their search mechanisms, although they were based on the same decoding method.