diff --git a/src/pages/index.vue b/src/pages/index.vue
index 4cb5eb0..e435194 100644
--- a/src/pages/index.vue
+++ b/src/pages/index.vue
@@ -899,8 +899,9 @@ export default {
           ),
         }));
 
+      const maxColumns = Math.min(3, Math.floor(window.innerWidth / 300));
       const result = this.dynamicSort
-        ? this.optimizeGridLayout(items)
+        ? items.sort((a, b) => a.order - b.order)
         : items.sort((a, b) => a.order - b.order);
 
       this.updateSortedItemsCache(key, result);
@@ -1712,56 +1713,24 @@ export default {
 
       try {
         const selectedDate = this.ensureDate(newDate);
-        const formattedDate = this.formatDate(selectedDate);
+        const dateStr = this.formatDate(selectedDate);
 
-        if (this.state.dateString !== formattedDate) {
-          this.state.dateString = formattedDate;
-          this.state.selectedDate = formattedDate;
-          this.state.selectedDateObj = selectedDate;
-          this.state.isToday =
-            formattedDate === this.formatDate(this.getToday());
+        if (dateStr === this.state.dateString) return;
 
-          this.$router
-            .replace({
-              query: { date: formattedDate },
-            })
-            .catch(() => {});
+        this.state.dateString = dateStr;
+        this.state.selectedDate = dateStr;
+        this.state.selectedDateObj = selectedDate;
+        this.state.isToday =
+          dateStr === this.formatDate(this.getToday());
 
-          // Load both data and subjects in parallel, force clear data when switching dates
-          await Promise.all([this.downloadData(true), this.loadSubjects()]);
-        }
+        // Load both data and subjects in parallel, force clear data when switching dates
+        await Promise.all([this.downloadData(true), this.loadSubjects()]);
       } catch (error) {
         console.error("Date processing error:", error);
         this.$message.error("日期处理错误", "请重新选择日期");
       }
     },
 
-    optimizeGridLayout(items) {
-      const maxColumns = Math.min(3, Math.floor(window.innerWidth / 300));
-      if (maxColumns <= 1) return items;
-
-      const columns = Array.from({ length: maxColumns }, () => ({
-        height: 0,
-        items: [],
-      }));
-
-      items.forEach((item) => {
-        const shortestColumn = columns.reduce(
-          (min, col, i) => (col.height < columns[min].height ? i : min),
-          0
-        );
-        columns[shortestColumn].items.push(item);
-        columns[shortestColumn].height += item.rowSpan;
-      });
-
-      return columns
-        .flatMap((col) => col.items)
-        .map((item, index) => ({
-          ...item,
-          order: index,
-        }));
-    },
-
     // 实时频道：加入设备房间并监听键变化
     setupRealtimeChannel() {
       try {
diff --git a/src/utils/gridLayout.js b/src/utils/gridLayout.js
new file mode 100644
index 0000000..a84af1c
--- /dev/null
+++ b/src/utils/gridLayout.js
@@ -0,0 +1,181 @@
+/**
+ * 优化网格布局算法
+ * 目标：使各列高度尽可能平均且最大高度最小
+ * 策略：LPT (Longest Processing Time) + 局部搜索 (Local Search)
+ *
+ * @param {Array} items - 待排序的卡片项，每项需包含 rowSpan 属性
+ * @param {number} maxColumns - 最大列数
+ * @returns {Array} - 排序后的卡片项，包含 order 属性
+ */
+export function optimizeGridLayout(items, maxColumns) {
+  if (maxColumns <= 1 || !items || items.length === 0) return items;
+
+  // 1. 初始分配：LPT (Longest Processing Time) 算法
+  // 按高度降序排序，优先处理大卡片
+  // 使用浅拷贝避免修改原数组
+  const sortedByHeight = [...items].sort((a, b) => b.rowSpan - a.rowSpan);
+
+  // 初始化列状态
+  // 使用 Int32Array 存储高度以提高性能（假设高度不会溢出）
+  const columnHeights = new Int32Array(maxColumns);
+  const columnItems = Array.from({ length: maxColumns }, () => []);
+
+  // 贪心分配
+  for (let i = 0; i < sortedByHeight.length; i++) {
+    const item = sortedByHeight[i];
+    // 寻找当前最矮的列
+    let shortestColIndex = 0;
+    let minHeight = columnHeights[0];
+
+    for (let j = 1; j < maxColumns; j++) {
+      if (columnHeights[j] < minHeight) {
+        minHeight = columnHeights[j];
+        shortestColIndex = j;
+      }
+    }
+
+    columnItems[shortestColIndex].push(item);
+    columnHeights[shortestColIndex] += item.rowSpan;
+  }
+
+  // 2. 优化阶段：尝试平衡最高和最低列
+  // 限制迭代次数，防止耗时过长
+  const MAX_ITERATIONS = 50;
+
+  for (let iter = 0; iter < MAX_ITERATIONS; iter++) {
+    // 找到最高和最低的列
+    let minIdx = 0;
+    let maxIdx = 0;
+    let minH = columnHeights[0];
+    let maxH = columnHeights[0];
+
+    for (let i = 1; i < maxColumns; i++) {
+      const h = columnHeights[i];
+      if (h < minH) {
+        minH = h;
+        minIdx = i;
+      } else if (h > maxH) {
+        maxH = h;
+        maxIdx = i;
+      }
+    }
+
+    const heightDiff = maxH - minH;
+    // 如果高度差很小，或者只有一列（逻辑上不可能，前面已拦截），则停止
+    if (heightDiff <= 1) break;
+
+    let bestAction = null;
+    let bestDiffReduction = 0;
+
+    const maxColItems = columnItems[maxIdx];
+    const minColItems = columnItems[minIdx];
+
+    // 策略 A: 尝试从高列移动一个卡片到低列
+    // 只需要检查能减少高度差的卡片
+    // 移动卡片 h，新高度差为 |(maxH - h) - (minH + h)| = |maxH - minH - 2h|
+    // 我们希望 |maxH - minH - 2h| < maxH - minH
+    for (let i = 0; i < maxColItems.length; i++) {
+      const item = maxColItems[i];
+      const h = item.rowSpan;
+
+      // 如果卡片高度大于高度差的一半，移动后反而可能导致低列变得比高列还高很多，需要检查绝对值
+      // 优化目标是最小化新的 max(newMaxH, newMinH) - min(newMaxH, newMinH)
+      // 但这里简化为只关注这两列的平衡
+
+      const newMaxH = maxH - h;
+      const newMinH = minH + h;
+      const newDiff = Math.abs(newMaxH - newMinH);
+
+      if (newDiff < heightDiff) {
+        const reduction = heightDiff - newDiff;
+        if (reduction > bestDiffReduction) {
+          bestDiffReduction = reduction;
+          bestAction = { type: "move", itemIdx: i, reduction };
+
+          // 如果已经找到非常好的移动（几乎完美平衡），可以提前结束搜索
+          if (newDiff <= 1) break;
+        }
+      }
+    }
+
+    // 策略 B: 尝试交换高列的一个大卡片和低列的一个小卡片
+    // 仅当策略 A 没有找到完美解时尝试
+    if (!bestAction || bestAction.reduction < heightDiff * 0.5) {
+      for (let i = 0; i < maxColItems.length; i++) {
+        const itemA = maxColItems[i];
+        for (let j = 0; j < minColItems.length; j++) {
+          const itemB = minColItems[j];
+
+          const hA = itemA.rowSpan;
+          const hB = itemB.rowSpan;
+
+          // 必须是高列拿出更大的卡片
+          if (hA <= hB) continue;
+
+          const change = hA - hB;
+          const newMaxH = maxH - change;
+          const newMinH = minH + change;
+          const newDiff = Math.abs(newMaxH - newMinH);
+
+          if (newDiff < heightDiff) {
+            const reduction = heightDiff - newDiff;
+            if (reduction > bestDiffReduction) {
+              bestDiffReduction = reduction;
+              bestAction = { type: "swap", idxA: i, idxB: j };
+            }
+          }
+        }
+      }
+    }
+
+    if (bestAction) {
+      if (bestAction.type === "move") {
+        const item = maxColItems[bestAction.itemIdx];
+        // 移除
+        maxColItems.splice(bestAction.itemIdx, 1);
+        // 添加
+        minColItems.push(item);
+        // 更新高度
+        columnHeights[maxIdx] -= item.rowSpan;
+        columnHeights[minIdx] += item.rowSpan;
+      } else {
+        const itemA = maxColItems[bestAction.idxA];
+        const itemB = minColItems[bestAction.idxB];
+        // 交换
+        maxColItems[bestAction.idxA] = itemB;
+        minColItems[bestAction.idxB] = itemA;
+        // 更新高度
+        const diff = itemA.rowSpan - itemB.rowSpan;
+        columnHeights[maxIdx] -= diff;
+        columnHeights[minIdx] += diff;
+      }
+    } else {
+      // 无法进一步优化
+      break;
+    }
+  }
+
+  // 3. 保持列内科目顺序并展平
+  // 预先计算总长度以分配数组
+  const result = new Array(items.length);
+  let resultIdx = 0;
+
+  for (let i = 0; i < maxColumns; i++) {
+    const colItems = columnItems[i];
+    // 列内排序
+    if (colItems.length > 1) {
+      colItems.sort((a, b) => a.order - b.order);
+    }
+
+    for (let j = 0; j < colItems.length; j++) {
+      // 复制对象以避免修改原始引用（如果需要纯函数特性）
+      // 这里为了性能直接修改或浅拷贝，根据需求调整
+      // 题目要求返回带 order 的新对象
+      const item = colItems[j];
+      result[resultIdx] = { ...item, order: resultIdx };
+      resultIdx++;
+    }
+  }
+
+  return result;
+}