为什么会收到： 'addition': 未连接的输入。每个层​输入必须连接到另一个​层的输出。

请问一下，我想使用深度网络设计器构建决斗结构的网络，并试图通过trainNetwork进行训练，为何使用additionlayer时会出现： 'addition': 未连接的输入。每个层输入必须连接到另一个层的输出。

以下为错误提示

以下是我的网络结构图。

以下是我的网络代码：

dqn_network = layerGraph();
tempLayers  = [
             featureInputLayer(32,"Name","current_state")
             fullyConnectedLayer(128,"Name","fc1")
             eluLayer(1,"Name","elu1")
             fullyConnectedLayer(128,"Name","fc2")
             eluLayer(1,"Name","elu2")
             fullyConnectedLayer(128,"Name","fc3")
             eluLayer(1,"Name","elu3")
             fullyConnectedLayer(128,"Name","fc4")
             eluLayer(1,"Name","elu4")
             fullyConnectedLayer(11,"Name","fc5")
             eluLayer(1,"Name","elu5")
             ];
dqn_network = addLayers(dqn_network,tempLayers);

tempLayers = [
    fullyConnectedLayer(128,"Name","fc")
    fullyConnectedLayer(11,"Name","fc_2")];
dqn_network = addLayers(dqn_network,tempLayers);

tempLayers = [
    fullyConnectedLayer(128,"Name","fc_1")
    fullyConnectedLayer(11,"Name","fc_3")];
dqn_network = addLayers(dqn_network,tempLayers);

tempLayers = [
    additionLayer(2,"Name","addition")
    fullyConnectedLayer(11,"Name","q_value")
    regressionLayer("Name","regressionoutput")];
dqn_network = addLayers(dqn_network,tempLayers);
% 清理辅助变量
clear tempLayers;
dqn_network = connectLayers(dqn_network,"elu5","fc");
dqn_network = connectLayers(dqn_network,"elu5","fc_1");
dqn_network = connectLayers(dqn_network,"fc_2","addition/in2");
dqn_network = connectLayers(dqn_network,"fc_3","addition/in1");