以限制規劃構建全年無休服務人員排班模式-以客服人員排班為例

Abstract

隨著經濟發展，人力成本亦不斷提高，對若干無法以科技自動化取代人工的全年無休服務專業而言，人員排班確是一項重要的課題。除了成本的考慮，排班公平性的良窳亦直接影響服務人員的士氣，進而衝擊到對顧客的服務品質。本研究以客服中心為例，來探討全年無休服務人員的排班問題。客服中心是當今電子商務時代，各大企業在網路上提供一週7天24小時服務的典型範例。客服中心一方面要滿足服務水準之人力需求，及遵守勞基法規、企業規定相關的各種硬限制條件，另一方面也要儘量滿足個別人員公平性的各種軟限制條件。再加上客服人員可能擁有雙重或多重技能使得客服中心人員排班的問題，遠比一般以人力成本最小化的排班問題更為複雜。 本研究擬將客服人員排班問題定式為一個限制滿足問題(CSP)，將之構建成限制規劃(CP)模式進行求解。而為有效率求解大規模的實務問題，本研究將問題分為二階段求解，第一階段為「排班」模式，產生滿足各條件限制及符合人力需求之班別及其數量，並考慮將不滿足需求程度及人力成本最小化。第二階段為「派遣」模式，為將第一階段產生之班表指派給人員，除考慮各項限制條件外亦考慮班表之公平性。 本研究以國內C銀行客服中心為個案，依據其實際排班相關規則及一業務組45名人員之排班資料作為測試例題，以ILOG OPL Studio軟體撰寫程式，於AMD 1.6G之個人電腦求解一個月之人員班表。測試結果可於10分鐘內完成整月班表求解，得到符合各項限制、97%的人力需求、人力成本最小之班表；在公平性方面，可達到當月各人員工作天數差異至多一日、假日休假天數差異至多二日。除班表外，本研究結果亦可提供人力供需統計與人員勞逸分配等資訊，供客服中心人力需求規劃及管理決策之依據。

For most service industries, the service agent scheduling or rostering is a major concern of the management because the high cost of service professionals can hardly be replaced by automated technology. In addition, the morale of the service crew highly relies on a fair roster or schedule, and this morale would easily translate into customer satisfaction, which in general represents the goal of a service business. Call center service agent scheduling problem is a typical case of 7-days-24-hours service crew scheduling problems. The scheduler not only has to provide a roster to satisfy all labor regulations and the service requirements (Hard Constraints) but also needs to consider to fit individual equitableness (Soft Constraints) as much as possible. The multiple-skill consideration, i.e. a single service agent can serve for multiple skill requirements at one time, makes the call center scheduling problem more complicated than general scheduling problem. In this research, we formulate the call center service agent scheduling problem as a constraint satisfaction problem (CSP), and using constraint programming (CP) method to solve the problem in two phases. The first phase is “Crew Scheduling”, which generates the optimal shift combinations to meet all work requirements and minimize demand unfitness and over-time labor cost. The second phase is “Crew Rostering”, which then generates equitable roster for each service agent to meet all related regulations. We conducted a case study of the call center of a Taiwan’s major bank. Using real world data of a division with 45 agents, we implemented the CP model on an AMD 1.6G personal computer. The ILOG OPL Studio served as the tool to construct our CP model. The results of a full-month roster of the case problem can be generated in 10 minutes. The results were checked by the case company, and obtained very positive response. In addition to the roster schedule results, our model also provides many useful information for workforce and manpower planning and management for a call center.