Violation Learning Differential Evolution-Based hp-Adaptive Pseudospectral Method for Trajectory Optimization of Space Maneuver Vehicle

Abstract

The sensitivity of the initial guess in terms of optimizer based on an hp-adaptive pseudospectral method for solving a space maneuver vehicle's (SMV) trajectory optimization problem has long been recognized as a difficult problem. Because of the sensitivity with regard to the initial guess, it may cost the solver a large amount of time to do the Newton iteration and get the optimal solution or even the local optimal solution. In this paper, to provide the optimizer a better initial guess and solve the SMV trajectory optimization problem, an initial guess generator using a violation learning differential evolution algorithm is introduced. A new constraint-handling strategy without using penalty function is presented to modify the fitness values so that the performance of each candidate can be generalized. In addition, a learning strategy is designed to add diversity for the population in order to improve the convergency speed and avoid local optima. Several simulation results are conducted by using the combination algorithm; simulation results indicated that using limited computational efforts, the method proposed to generate initial guess can have better performance in terms of convergence ability and convergence speed compared with other approaches. By using the initial guess, the combinational method can also enhance the quality of the solution and reduce the number of Newton iteration and computational time. Therefore, the method is potentially feasible for solving the SMV trajectory optimization problem.

Multiobjective Overtaking Maneuver Planning for Autonomous Ground Vehicles

“…Based on the generated pareto front, two extreme solutions, along with one compromised solution, can be detected (as indicated 8 in Fig.5). The objective and the constraint violation values of these three solutions are tabulated in Table IV. From Table IV, extreme point 1 can be treated as the solution with the best overtaking visibility but the worst overtaking time and path smoothness.…”

“…Here, c denotes the number of constraints defined in (8). The way to calculate i( j ) for the constraint i(·) ≤ * i can be written as:…”

“…One important advantage of using this technique is that it has the capability to model different mission-dependent requirements into constraints, which can then be entailed in the optimization formulation. Contributions made to design or apply optimizationbased motion planners are available in the literature [8]- [11]. For instance, in [8], a space maneuver vehicle regional observation trajectory was designed by applying a desensitized motion optimizer.…”

“…Contributions made to design or apply optimizationbased motion planners are available in the literature [8]- [11]. For instance, in [8], a space maneuver vehicle regional observation trajectory was designed by applying a desensitized motion optimizer. In [10] the authors applied an interior-point method to evaluate the time-optimal trajectory of an automatic parking problem.…”

“…III. FUZZY ADAPTIVE MULTI-OBJECTIVE PARTICLE SWARM OPTIMIZATION ALGORITHM In this section, an optimization-based maneuver planning approach is presented to solve the problem defined by (8). This approach, named fuzzy adaptive multi-objective particle swarm optimization (FAMOPSO), can be treated as an enhanced version of the PSO-based motion planner reported in [24].…”