為了配合行政院「生物技術產業推動方案」中的「健全臨床試驗體系」，
行政院正積極輔導醫院設立臨床試驗管理機構，
透過試驗醫院管理機構(Site Management Organiza- tion，SMO)之
管理模式取代早期單一試驗案由單一位醫師負責的管理模式，
以有效管理研究護士及病患間的臨床試驗問題，並達到擴大病患母體範圍，縮短收案時間的目的。
早期管理模式由於病患來源分散，需要較長的收案時間；而試驗醫院管理機構(SMO)之
管理模式有較複雜的護士服務病患規定，
因此如何以較少的人力資源於將所需的病患個數收齊，為本論文的研究主軸。
根據研究調查，試驗醫院管理機構之管理模式應用於人體試驗流程的第三階段，而此階段所耗費的成本相當大。
為了使其管理模式除了擁有縮短收案時間的優勢，同時能夠根據其管理方式來減少人事成本，
本研究根據其管理流程，建構最佳化數學規劃模式，
並以該模式為基礎提供較好的療程與人事規劃以達成降低人事成本的目的。

此數學模式與傳統的護士人員排程問題(Nurse Rostering Problem，NRP)
類似，在數學規劃領域中皆屬於整數規劃的問題。 透過參訪
使用SMO管理模式的公司，本研究將計劃試驗案中護士與病患間的療程規劃寫成數學模式，
並利用CPLEX軟體求解。然而，由於所求得之CPLEX最佳排程不甚人性化，
因此本研究亦發展更有效率的啟發式解法以在更短的時間內提出一份好的療程規劃，
並以該啟發式解法為核心來建置一套人力規劃系統，以提供公司管理者決策之參考。

In order to carry out the clinical test system of the
biologicaltechnology industry plan, the Executive Yuan has been
actively helpinghospitals establish clinical trial management.
Site ManagementOrganization (SMO) has replaced the old management
model that onedoctor handles one plan at a time. SMO can
effectively help schedulenurses to serve more patients and reduce
the time to search forpatients. The original management model
before SMO usually requireslonger time to search for patients
because the sources of the patientsare scattered around places. On
the other hand, SMO has more complexrules to assign nurses to the
patients. Our research focuses ontechniques to search for and also
serve all patients with less humanresource. According to our
studies, SMO is used in the third stage ofthe clinical trial
process which is also the stage of the largestcost. In order to
improve the operations of SMO, one needs toefficiently decide the
schedules of treatments for each patient andeach served nurse so
that the cost for human resources is reduced.This optimization
problem can be considered as a Nurse RosteringProblem (NRP), which
is often solved by integer programming.In this thesis, we propose
a mathematical model based on integerprogramming that leads to
better schedules and reduces the cost ofhuman resource. In
particular, several objectives including theminimization of the
number of served nurses, and balancing the dailyworkload are
formulated in our IP models and solve by CPLEX. However,the
solution calculated by CPLEX seems to be inconvenient comparedwith
the original schedule in the sense that the solution tends
topartition the daily schedule for each nurse. Moreover, CPLEX
consumesa lot of computational time. To construct a more
convenient schedulein a shorter time such that the original
objectives are achieved asmuch as possible, we propose several
greedy heuristics and build arostering information system to help
SMO managers design a goodschedule that not only reduces the human
resource costs but alsoassigns workload more reasonably.

Aickelin, U. 1999. Genetic algorithms for multiple-choice problems. Unpublisheddoctoral dissertation, University of Wales, Swansea.

Aickelin, U. and Dowsland, K. Exploiting problem structure in a genetic algorithm approach to a nurse rostering problem. Journal of scheduling, 3, 139-153, 2000.

Baker, K. R. Workforce allocation in cyclical scheduling problem: A survey*. Operational Research Quarterly, 27(1), 155{167, 1976.

Bard, J. F. and Wan, L. The task assignment problem for unrestricted movementbetween workstation groups. Journal of Scheduling, 9(4), 315{341, 2006.

Barnhart, C., Johnson, E., Nemhauser, G., M., S. and Vance, P. Branch-and price:Column generation for solving huge integer programs. Operation Research, 46(3),316-329, 1998.

Brusco, M. and Jacobs, L. A simulated annealing approach to the solution of °exiblelabor scheduling problems. Journal of the Operational Research Society, 44(12),
1191-1200, 1993.

Burns, A., Hayes, N. and Richardson, M. Generating feasible cyclic schedules. Control Engineering Practice, 3(2), 151-162, 1995.

Cai, X. and Li, K. A genetic algorithm for scheduling sta® of mixed skills inder multicriteria. European Journal of Operational Research, 125, 359-369, 2000.

Dodin, B., Elimam, A. and Rolland, E. Tabu search in audit scheduling. European Journal of Operational Research, 106, 373-392, 1998.

Dowling, D., Mackenzie, H., Krishnamoorthy, M. and Sier, D. Sta® rostering at a largeinternational airport. Annals of Operational Research, 125-147, 1997.

Dowsland, K. Nurse scheduling eith tabu search and strategic oscillation. European Journal of Operational Research, 106, 393-407, 1998.

Emden-Weinert, T. and Proksch, M. Best practice simulated annealing for the airline crew scheduling problem. Journal of Heuristics, 5, 419-436, 1999.

Ernst, A. T., Jiang, H., Krishnamoorthy, M. and Sier, D. Sta® scheduling and rostering:A review of applications, methods and models. European Journal of Operational
Research, 153, 3{27, 2004.

Heus, K. and Weil, G. 1996. Constriant programming a nurse scheduling application. In Proceedings of the second international conference on the practical application
of constraint technology (p. 115-127).

Hoffman, K. and Padberg, M. Solving airline crew scheduling problems by branch-and-cut. Management Science, 39(6), 1993.

Jaumard, B., Semet, F. and Vover, T. A generalized linear programming model for nurse scheduling. European Journal of Operational Research, 107(1), 1{18, 1998.

Kathryn, A. and Dowsland, K. Nurse scheduling with tabu search and strategic oscillation. European Journal of Operation Research, 106, 393-407, 1998.

L., M. J. State of the clinical trials industry (a sourcebook of charts and statistics).Thomson CenterWatch, 2005.

Lawler, E. W. and Wood, D. E. Branch-and-bound method@a survey. Operations Research, 14, 699-719, 1966.

Ryan, D. The solution of massive generlized set partitioning problems in aircrew rostering. Journal of the Operation Research Society, 43(5), 459-467, 1992.

Tackling, F. E. 1998. Scheduling problems using integer programming. Unpublished master's thesis, University of Wales, Swansea.

Tanomaru, J. 1995. Staff scheduling by a genetic algorithm with heuristic operators. In Proceeding of the 1995 ieee international conference inevolutionary computer
science (p. 456-461).

Thompson, G. A simulated-annealing heuristic for shift scheduling using noncontinuously available emplyoees. Computers and Operations Research, 23(3),275-288, 1996.

Wedelin, D. An algorithm for 0-1 programming with an application to airline crew scheduling. Annals Operation Research, 57, 283-301, 1995.