單場站公路客運司機員與車輛排班問題之研究

SINGLE-DEPOT BUS DRIVERS AND VEHICLES SCHEDULING PROBLEM

公路客運司機員與車輛排班乃是指公路客運營運機構，依據營運路線及班次時刻表，所進行司機員與車輛之勤務規劃作業，為公路客運營運作業中之重要作業項目。本研究考量實務作業特性，從經營者營運目標與司機員需求兩層面構建數學規劃模式，並發展以勤務為染色體之基因演算法進行模式求解，該演算法透過尋偶式交配法與染色體突變法的提出，改善傳統基因演算法容易造成班次重複或佚失問題之缺失。經實例測試結果，顯示本研究除可兼顧求解時間與求解品質外，亦可獲得較目前營運班表為佳之分析結果。

This paper applies a genetic algorithm with a new method which canexecute without trips over-cover or uncover in the process of crossover andmutation on the problem “bus drivers and vehicles scheduling”. For most buscompanies that operate on a fixed timetable, this problem should be consideredaccording to both drivers’ and operators’ demand and constraints by laws orsome other rules. In this algorithm, we use a weighted Hölder norm in thefitness function to combine different objectives. After applying the new methodsto some real cases, it is proved to be able to reduce human effort and obtain afeasible and better schedule.