摘要: | 警察為維護社會治安與交通安全為主要任務之角色,其重要性不可言喻。警政署統計資料指出,統計至2018年2月底為止,全國實際警察預算缺額仍有3398人,分局各派出所基層警力往往須透過加班,以因應人力短缺之現象。然而,實務上分局各派出所之班表主要依決策者人工經驗進行編排,除缺乏效率外,亦難以掌握規劃品質。此外,於警察人力供給規劃上,分局各派出所目前皆獨力編排班表,較缺乏考量相互支援策略以進行規劃。因此,於目前警察人力不足之環境下,如何有效運用分局各派所警察人力仍為當前相當重要課題之一。 本研究透過數學規劃方法,以分局各派出所所長為立場,以警察人力供給小時最小化為目標,考量不同供給策略、實務狀況等限制,構建確定性與隨機性需求下例行性勤務分局各派出所警察人力供給規劃之獨力調派模式及相互支援模式,期望規劃結果能輔助警政相關單位有效地編排班表。求解方法方面,確定性獨立調派模式、確定性相互支援模式及隨機性獨立調派模式皆直接透過數學規劃軟體CPLEX配合C++程式語言進行求解,至於隨機性相互支援模式,其問題規模龐大,加上涉及整數與實數計算問題,導致無法有效率透過數學規劃軟體求解,故本研究發展一啟發式演算法有效地進行求解。另外,本研究利用隨機模式評估指標以評估隨機模式規劃效益。本研究針對國內某分局各派出所人力供給相關資料及合理假設之輸入資料,進行範例測試。測試結果顯示本研究模式規劃效益明顯優於實務編排方式,並且隨機性規劃及啟發解法應用績效皆表現良好。除此之外,相互支援模式之規劃效益較獨力調派模式優。本研究期望所發展的模式及求解方法除能提供學術界參考之外,其規劃結果更能提供警政相關單位進行最佳警察人力供給規劃,幫助警政相關單位提升人力供給配置之效力,減少人力不足以致過勞問題,進而提升警政相關單位在警力資源管理上之績效。 ;The primary duty of a police officer is to maintain public security and traffic safety. According to the National Police Agency, there are still 3398 vacant positions until the end of February 2018. therefore, most police officers need to work overtime. However, the police officer manpower supply plans for police stations in Taiwan are typically generated with decision maker experiences currently, which are neither efficient nor effective. It is difficult to control the planning quality, usually resulting in inadequate manpower supplies. In addition, each police station of a precinct in Taiwan generally performs the manpower supply planning without considering mutual support strategy. Hence, how to efficiently use their manpower under scarce human resources is a majar issue for police divisions/departments. ln this research, based on the police division’s point of view, we adopt the mathematical programming method to construct deterministic-demand and stochastic-demand routine police officer manpower supply models, totally including 4 models, by considering different supply strategies and related operating constraints, with the objective of minimizing total manpower supply-hours. We expect that the results can assist police stations to plan their shift schedule. We utilize C++ computer language, coupled with the CPLEX, a mathematical programming solver, to solve the deterministic models and the stochastic independent assignment model. As to the stochastic mutual support model, since its’ problem size is too huge to be efficiently solved by the CPLEX. Therefore, we develop the heuristics according to the problem characteristics. Besides, we utilize EVPI and VSS to evaluate the performance of the stochastic models. The numerical tests using the data collected from Taiwan police department/divisions are performed. The results show that the police officer manpower supply plan by applying the models is better than that by pragmatic decisions, and the performance of the stochastic models and the heuristics are both good. Besides, the performance of a mutual support model is better than an indedendent assignment model. Finally, these models with the solution methods are expected to be useful not only for academics, but for police stations to generate optimal manpower supply plans, in order to reduce the shortage of police officers and improve the performance in managing the police officer human resources. |