Α two-phase adaptive variable neighborhood approach for nurse rostering

Genetic based discrete particle swarm optimization for Elderly Day Care Center timetabling

“…Nurse rostering problems generate a schedule for each nurse, who has day off patterns, working shifts patterns and different work contracts, to fulfill the collective agreement requirements and hospital staffing demand coverage, while minimizing salary cost and maximizing nurse preferences and quality [3][4][5][6][7]. Burke and Curtois [6] developed a mathematical model for all the instances of nurse rostering problems by applying "regular expression" to incorporate their varying types of constraints (e.g., minimum/maximum consecutive work days, day on/off request, and shift on/off request).…”

“…Burke and Curtois [6] developed a mathematical model for all the instances of nurse rostering problems by applying "regular expression" to incorporate their varying types of constraints (e.g., minimum/maximum consecutive work days, day on/off request, and shift on/off request). Solos et al [3] proposed a more effective generic variable neighborhood search algorithm to solve seven different nurse rostering instances and summarize these varying types of constraints into two categories: hard constraints (e.g., all shift type demands must be met) and soft constraints (e.g., maximum number of hours worked), most of which were also modelled as an integer programming in [5] and included in [7] when presenting a mathematical formulation for all nurse rostering problem instances with 2 hard and 18 soft constraints in the First International Nurse Rostering Competition (INRC-2010). It is different from the educational timetabling problem mainly because of the demand coverage constraint, which specifies the number of nurses in each skill level [4], salary costs, nurse preferences, and degree of balance among nurses.…”

“…The EDCC timetabling aims to find a feasible solution with smallest objective value and determine which soft constraints suffer the most violation. Timetabling problems are encountered in various situations, such as rostering duties of nurse in hospitals [3][4][5][6][7] , scheduling transportation events [8], and constructing timetables for courses or examinations in the education industry [9][10][11][12][13][14][15]. The EDCC timetabling has some unique characteristics (e.g.…”

A hybrid variable neighbourhood search and dynamic programming approach for the nurse rostering problem

“…Indeed, almost all NRPs in reality imply such succession constraints. Therefore, due to the computational complexity of the problem, exact solution methods mostly fail to solve large size instances in a reasonable time [33], [28]. Therefore, heuristic solution methods are adopted to obtain good solutions within reasonable computation time.…”

“…Their algorithm was able to provide new upper bounds for 12 instances and matching the best known results for 39 instances out of 60 instances. A two-stage VNS algorithm for the NRP was proposed in [33], [32], where the first stage deals with the assignment of nurses to working days, then the second stage deals with the assignment of nurses to shift types. Experimental results showed that, in addition to the superiority of the two-stage VNS algorithm compared to six other effective algorithms [32], very comparable results with the five finalists of the First International Nurse Rostering Competition were reported [33].…”