MOSQP is a multiobjective optimization solver for bound, linear, and nonlinear constrained problems (for which the derivatives of the objective functions and constraints exist).
The algorithm performs Sequential Quadratic Programming type iterations to build an approximation to the Pareto front.
MOSQP makes use of derivative information of the objective functions and constraints to build quadratic models from which search directions are obtained. It has been shown to be efficient and robust for a large set of bound and constrained multiobjective optimization problems.
The code is written in MATLAB. It provides an interface with AMPL.
Benchmarks (MOSQP vs NSGA-II)
Benchmarks (MOSQP vs NSGA-II vs MOScalar)
Pareto fronts (MOSQP vs MOScalar and MOSQP vs NSGA-II)
Downloading and Conditions for Use
Conditions for Use:
MOSQP is freely available for educational, research, and commercial use, but we expect that all publications describing work using this software quote the reference given below.
J. Fliege and A. I. F. Vaz, A method for constrained multiobjective optimization based on SQP techniques, SIAM J. Optim., 26(4), 2091–2119.
Download (version 1.1):