This paper discusses the worst-case tolerance optimization of linear arrays

using excitation coefficients as design parameters.The problem considered is defined mathematically as a worst-case problem

a fixed tolerance problem（FTP）and a variable tolerance problem（VTP） and the new algorithms for solving this problem is presented.For solution of the FTP

we use the regular polyhedron method.The VTP is solved by a double iterative algorithm in which the inner iteration is performed by the FTP algorithm and the outer iteration is implemented by the comparison method.The VTP

which uses the largest discrepancy between the desired and the actual patterns in the tolerance region as the standard

harmonizes the relation between the array sklelobe level and the tolerance of excitation coefficients

so that the sidelobe level satisfies the desired requirements and the tolerance region is as large as possible.Finally

the numerical results of the 6-element and 16-element linear array tolerance optimization designs are given.