本研究探討保全公司運鈔車路線與時間排程規劃，近十年來運鈔車路線多由人工方式安排，每日路線單調缺乏變化性，因此常常被歹徒預測並下手搶劫，這十年間台灣搶奪運鈔車竟高達35 件之多。過去已有研究以數學規劃軟體CPLEX 分為三階段求解排程規劃，但由於其分為三階段求解，較無法考慮到近似全域最佳解的情形，因此本研究以基因演算法為基礎，設計求運鈔車之排程規劃，以求得近似最佳排程。
除了考慮到時間變化性與空間變化性之外，本研究將排程變化差異程度隨機化，如此一來，讓歹徒無法從過去排程的趨勢來推測未來的路線安排，在時間差異與空間差異的限制條件下，以總成本最小化為目標。利用基因演算法之選擇、交配、突變來演化出近似最佳解之排程規劃。由小範例顯示，基因演算法求得的近似解非常逼近窮舉法所求得的最佳解，表示本研究所提出的方法應該可以應用於運鈔車排程規劃；並將基因演算法應用到更大的範例，基因演算法所得結果遠優於隨機產生的排程規劃結果。

This paper discusses the planning of armored car route and time schedule for security service firm. Over the past decade, most armored car routes are arranged manually so that monotonous daily route lacks of variation. These routes were often predicted by criminals and they proceeded with robbery. In the past decades, the robbing incidents in Taiwan have reached up to 35 cases. Although the approximate solution is solved by CPLEX, it is divided into three stages so that it is unable to take into account the approximate global optimum solution. Therefore, this paper develops a heuristic based on genetic algorithm to solve the armored car route scheduling in order to achieve the near-optimal solution.
In addition to considering the time variability and spatial variability, this study randomize the variety difference. As a result, criminals are unable to infer the routing in the future from the past trend of scheduling. Under the restricted conditions of time differences and spatial differences, the goal is to minimize the total cost. Through the selection, crossover and mutations, a near-optimal solution of the route scheduling is found. By a small example, it shows that the approximate solution obtained by the genetic algorithm is very near the optimal solution, thus proving that the proposed method in this study could be applied to the armored car route planning. By applying to a larger example, genetic algorithms obtained results far superior to randomly generated schedule planning results. Finally, this paper presents recommendations and conclusions.