The Study On The Course Scheduling Problem Of Multi-objective And Constraint Condition Based On The Improved Genetic Algorithm

With the implementation of policy in the college expansion plan,the school-running scale in each college is gradually increasing,so that the traditional manual scheduling assignment can't meet the requirements for normal teaching work in colleges.On the one hand,considering that scheduling work gets involved in more factors and there are mutual association and influence in each factor,the solution difficulty of the scheduling problem becomes even stronger.On the other hand,the evaluation index with the student first,teacher-orientated,full utilization and rational distribution of resources are also more complicated with the in-depth higher education study.As a result,the scales and difficulties in the course scheduling problem of each college are gradually increasing.The necessity of greatly solving the scheduling problem is standing out day by day.The essence of the scheduling problem means to allocate reasonable and conflict-free teaching resources to each teaching task as satisfying the specific constraint conditions in the scheduling process.It is a typical multi-objective and multi-constraint combination problem.For this reason,on the basis of the combining with the genetic algorithm,the author fully analyzed the features of the scheduling problem and carried out relevant studies on each constraint condition,conflict detection and elimination,solution,and optimization in the scheduling problem.The study is stated as following :In order to gain a higher quality for the scheduling problem,the author fully analyzed the scheduling principles,constructs the multi-objective scheduling problem model from the perspectives of schools,teachers and students,and states the necessity one by one.The multiobjective optimization model designed for the scheduling principles can perfectly described the solution goals to be considered in the scheduling problem of colleges.The basic precondition of the reasonable scheduling is the satisfaction of the constraint conditions.For this reason,the author firstly integrated with the scheduling constraint conditions,defined the necessity degree of each constraint condition,and designed multiple conflict detection algorithms in accordance with different constraint conditions.Considering that constraint conditions affect and restrain each other,the author elaborated the situation of new conflicts caused by eliminating some conflicts.Under the circumstance,the term-by-term searching algorithm for conflict resolution was designed.The experimental results proved that the above-mentioned algorithm can eliminate all conflicts in the scheduling problem and it had good practicability.By aiming at the main features in the scheduling problem,the author designed the improved genetic algorithm to solve the multi-objective optimization problem.First of all,the author put forward the variable-length decimalism coding scheme that can satisfy the same course to arrange in different time,different classrooms and different teaching weeks.This coding scheme gived full considerations to the flexibility of the course classroom and time arrangement,thus the scheduling problem has good rationality.At the same time,the author designed the local search operator by aiming at the specific problem and accelerated the convergence rate of the algorithm.At last,under the framework of reserving the optimal individuals,the author improved the selective operator,cross operator and mutation operator.The experiment proved that the designed algorithm not only has the fast convergence rate,but also can improve the individual diversity to some extent,so as to achieve goals of enhancing the search space and jumping out of the local optimum.