Periodic Mutation Operator for Nurse Scheduling by Using Cooperative GA

This paper proposes an effective mutation operator for Cooperative Genetic Algorithm (CGA) to be applied to a practical Nurse Scheduling Problem (NSP). NSP is a complex combinatorial optimizing problem for which many requirements must be considered. The changes of the shift schedule yields various problems, for example, a drop in the nursing level. The author describes a technique of the reoptimization of the nurse schedule in response to a change. CGA well suits local search, but its failure to handle global search leads to inferior solutions. CGA is superior in ability for local search by means of its crossover operator, but often stagnates at the global search. To solve this problem, a mutation operator activated is proposed depending on the optimization speed. This mutation yields small changes in the population depending on the optimization speed. Then the population is able to escape from a local minimum area by means of the mutation. However, this mutation operator is composed of two well-defined parameters. This means that users have to consider the value of the parameters carefully. To solve this problem, a periodic mutation operator is proposed which has only one parameter to define itself. This simplified mutation operator is effective over a wide range of the parameter value. DOI: 10.4018/jaec.2012070101 2 International Journal of Applied Evolutionary Computation, 3(3), 1-16, July-September 2012 Copyright © 2012, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. 2003; Cheang, Li, Lim, & Rodrigues, 2003; Burke, De Causmaecker, & Berge, 2004; Burke, De Causmaecker, Berghe, & Landeghem, 2004; Burke, 2001; Burke, De Causmaecker, Petrovic, & Berge, 2006; Ernst, Jiang, Krishnamoorthy, Owens, & Sier, 2004; Li & Aickelin, 2004; Bard & Purnomo, 2005, 2007; Oezcan, 2005). In an early study (Goto, Aze, Yamagishi, Hirota, & Fujii, 1993), NSP, defined as a discrete planning problem, is solved by using Hopfield model type neural network. Berrada, Ferland, and Michelon (1996) have proposed a technique to define the nurse scheduling problem as a multi-objective problem and to solve it by using a simple optimizing algorithm. The technique by Takaba, Maeda, and Sakaba (1998) provides a simple editing tool and simple GA for the nurse scheduling under Visual Basic environment. There are several techniques (Ikegami, 2001; Inoue, Furuhashi, Maeda, & Takabane, 2002; Bard & Purnomo, 2005, 2007) that require the user to modify or select the nurse schedule in the middle or the final stage of the optimization. Burke, De Causmaecker, and Berge (2004), Burke, De Causmaecker, Berghe, and Landeghem (2004), Burke (2001), and Burke, De Causmaecker, Petrovic, and Berge (2001, 2006) apply a memetic approach to the nurse scheduling problem. Burke, De Causmaecker, Petrovic, and Berghe (2001) also define a technique to evaluate the nurse schedule. Croce and Salassa (2010) propose a variable neighborhood search technique for the nurse scheduling. However, the scheduling problem defined in this manuscript is too easy. And the technique proposed in this manuscript is applied to a private hospital in Italy. Real problem of the nurse scheduling in the general hospital is not so easy and very hard to solve. Some of these techniques are implemented in commercial nurse scheduling software. However, the evaluation technique does not fit to the shift system of our country. Moreover, such a commercial software has not been utilized in most hospitals in fact, because a schedule given by the commercial software is unsatisfactory. The operation of the software is very complex for the user too. Therefore, we have defined the evaluation technique of the nurse schedule (Ohki, Morimoto, & Miyake, 2006; Ohki, Uneme, Hayashi, & Ohkita, 2007; Uneme, Kawano, & Ohki, 2008). In the real hospital, there are cases that nurses attend on a different day from the original schedule because of circumstances of other nurse or an emergency. There are also the cases that a nurse whom duty has been assigned originally takes a rest due to a disease. We discuss such a case that the nurse schedule has been changed in the past weeks of the current month. Such changes yield various inconveniences, for example, imbalance of the number of holidays and attendances. Such an inconvenience causes the fall of the nursing level of the whole nurse organization. Therefore, the inconvenience should be eliminated to make a better schedule. By considering the change of the shift schedule whenever one week passes, the shift schedule is reoptimized in remaining weeks of the current month. The shift schedule generated by the commercial software is unsatisfactory. And, many interactions to readjust the schedule are also very complex. In fact, the nurse schedule is still made by the hand of a manager or a chief nurse in many general hospitals. In our investigation, there are no general hospitals using the commercial software for nurse scheduling. The optimization algorithm of such the commercial software is still poor, and moreover, the schedule provided by such the software is hard to revise too. In this paper, we discuss on generation and optimization of the nurse schedule by using the Cooperative Genetic Algorithm (CGA) (Itoga, Taniguchi, Hoshino, & Kamei, 2003). CGA is a kind of Genetic Algorithm (GA) (Goldberg, 1989), and powerful optimizing algorithm for such a combinatorial optimization problem. In GA, Individuals compete with each other and superior individuals are preserved. On the other hand, individuals cooperate with each other and the optimization of whole population progresses in CGA. The conventional CGA optimizes the nurse schedule only by using a crossover operator, because the crossover has been considered as 14 more pages are available in the full version of this document, which may be purchased using the "Add to Cart" button on the product's webpage: www.igi-global.com/article/periodic-mutation-operator-nursescheduling/68830?camid=4v1 This title is available in InfoSci-Journals, InfoSci-Journal Disciplines Computer Science, Security, and Information Technology. Recommend this product to your librarian: www.igi-global.com/e-resources/libraryrecommendation/?id=2