Computed School Assignments in a Large District

The US Supreme Court decisions of 1954 on de jure segregation of schools, and recent decisions denying de facto segregation and barring “freedom-of-choice” pupil assignment, have forced school districts to devise plans for integrating public schools. Finding a feasible method of achieving racially balanced schools is difficult at best. A great number of factors must be taken into consideration. Recently, we have applied three linear programming models to a large school district (45 elementary schools, 20,000 students); they demonstrate the extent to which desegregation can be attained at moderate cost. The problems discussed in this paper include the collection and organization of data, and the problem formulation modifications necessary to conform to computer programs and storage limits. Some resulting assignment patterns are shown. Objective functions and reasonable constraints are suggested, but these can be modified. More important than the solution of a single problem is the fact that each solution represents a policy or point of view. School authorities, judges, and the public can see the logical implications of each point of view and select school assignment plans on a rational basis.