Hai-lin Liu

A Novel Weight Design in Multi-objective Evolutionary Algorithm

This paper presents a method to improve the performance of MOEA/D. The idea is to approximate the Pareto front(PF) by using a linear interpolation of the non-dominant solutions. It propose a novel weight design method for multi-objective evolutionary algorithm. Even when the PF is complex, we can obtain the Pareto optimal solutions which are distributed uniformly over the PF. Some test functions are constructed to compare the performance of the proposed algorithm with that of MOEA/D. The results indicate that the proposed algorithm could significantly outperform MOEA/D on these test instances.

Evolutionary Algorithm for the Radio Planning and Coverage Optimization of 3G Cellular Networks

WCDMA network planning decides on where to locate base stations (BS) and which configuration to select considering quality constraints and coverage constraints in order to maximize coverage and minimize costs. Classical approaches optimize these two objectives by combining them together in a single objective. In this paper a multi-objective evolutionary algorithm based on determined weight and sub-regional search is specialized to solve the problem. A specific encoding represents configuration parameters combination of a base station. Whether or not to coverage for each test point (TP) is determined by the location and the configuration of the BSs. We simplify the model by computing the transmitting power (mobile phone to base station) of mobile phone of each test point and the link power (base station to mobile phone). This reasonable simplicity is efficiently. A local search operator is designed for the multi-objective evolutionary algorithm. Simulation results have shown the effectiveness of the proposed evolutionary components by providing a high quality set of alternative solutions.

Multi-objective Evolutionary Algorithm for 4G Base Station Planning

Planning of base stations (BS) is one of the fundamental problems in the fourth generation (4G) wireless network design. A new mathematical model for the 4G BS planning is proposed in this paper. The co-channel interference, orthogonal frequency division multiplexing (OFDM), the cell edge rate, the reference signal received power (RSRP) and the base station density are considered in detail in this model. The objectives of the model are to minimum construction cost, maximum coverage and capacity at the same time. Its a multiobjective optimization problem with constraints. So a multiobjective algorithm with a local search operator is designed to solve this model. Simulation results show that the according algorithm with local search operator can give a set of solutions with relatively lower cost, larger coverage, larger capacity and faster speed than the algorithm without local search operator. The effective and efficient of the proposed model can also be identified by the comparison of the solutions.

A new fuzzy K-EVD orthogonal complement space clustering method

This paper studies the problem of underdetermined blind source separation with the nonstrictly sparse condition. Different from current approaches in literature, we propose a new and more effective algorithm to estimate the mixing matrices resulted from noise output data sets. After we introduce a clustering prototype of orthogonal complement space and give an extension of the normal vector clustering prototype, a new method combing the fuzzy clustering and eigenvalue decomposition technique to estimate the mixing matrix is presented in order to deal with the nonstrictly sparse situation. A convergent algorithm for estimating the mixing matrices is established, and numerical simulations are given to demonstrate the effectiveness of the proposed approach.

Resource Allocation for OFDM System Based on Evolutionary Algorithm

The conventional resource allocation method is so-called two-step method in OFDM system. The two-step method can reduce the computational complexity. However, its solution accuracy is not very good. This paper proposes an improved two-step method. This method combines evolutionary algorithm with simulated annealing thought, and can take into account the sub-carrier distribution and power distribution at the same time. Meanwhile this paper improves the evaluation criteria for fairness, so that the algorithm can better balance between capacity and fairness. Numerical experiments show that the proposed algorithm is effective.

Preference-Based Evolutionary Multi-objective Optimization

Evolutionary Multi-objective Optimization (EMO) approaches have been amply applied to find a representative set of Pareto-optimal solutions in the past decades. Although there are advantages of getting the range of each objective and the shape of the entire Pareto front for an adequate decision-making, the task of choosing a preferred set of Pareto-optimal solutions is also important. In this paper, we combine a preference-based strategy with an EMO methodology and demonstrate how, instead of one solution, a preferred set of solutions in the preferred range can be found. The basic idea is that each objective function corresponds to a marginal utility function, which indicates the decision-makers preferred range for each objective. The corresponding utility function denotes the decision-makers satisfaction. Such procedures will provide the decision-maker with a set of solutions near his preferred ranges so that a better and more reliable decision can be made.

A New Algorithm Estimating the Mixing Matrix for the Sparse Component Analysis

For the problem of Underdetermined Blind Source Separation (UBSS) under nonstrictly sparse condition, we propose a new algorithm to estimate mixing matrix. Firstly, a clustering prototype of orthogonal complement space is introduced; Secondly, a fuzzy EVD clustering method which combines fuzzy clustering and Eigenvalue Decomposition (EVD) is presented. Based on these two methods, the algorithm proposed in this paper is robust against noise without losing convergence speed.

Multi-Objective Evolutionary Algorithm Based on Dynamical Crossover and Mutation

In complicated multi-objective optimization, it often happens that points in part region of Pareto front are easy to get, but in others are difficult. To obtain evenly distributed Pareto optimal solution, we construct dynamical crossover and mutation probability which can self-adaptively adjust the number of individuals engaged in crossover and mutation, combine with the fitness function constructed by weighted min-max strategy in which the weight is uniformly designed, to present a new multi-objective evolutionary algorithm (DMOEA). To evaluate the performance of our algorithm, we compare the numerical results of our algorithm with the MOEA/D-DE and NSGA-II-DE, the comparison shows that our algorithm is very efficient.

A fast evolutionary algorithm with searching preference

Evolutionary algorithm EA has been successfully applied to many numerical optimisation problems in recent years. However, EA has rather slow convergence rates on some optimisation problems. In this paper, a fast evolutionary algorithm FEA with searching preference is proposed. Our basic idea is that the better an individual is, the more resources are invested to search the region close to the individual. Two techniques are applied to achieve it: making the best individual found so far always take part in crossover and mutation, and proposing a novel crossover and mutation operator based on simulated annealing. Obviously, the search process emphasises the region around the best individual. Furthermore, we can prove that FEA converges to a global optimum in probability. Numerical simulations are conducted for 19 standard test functions. The performance of FEA is compared with three EAs FEP, OGA/Q and LEA that have been published recently. The results indicate that FEA is effective and efficient. Furthermore, the result obtained by FEP is better than the best result found so far for ƒ

4G Heterogeneous Networks Base Station Planning Using Evolutionary Multi-objective Algorithm

Heterogeneous networks (Het Net) is the development trend of future wireless networks, but the additive small base stations (BSs) make existing BS planning models unavailable. This paper builds a new BS planning model in which takes signal coverage, system capacity and cost as objective functions and takes interference as a very important constraint. Since these issues are conflicting, the optimization problem is a multi-objective optimization problem with constraints which is difficult to tackle. Evolutionary multi-objective optimization algorithm is considered to solve the problem in this paper. In order to improve the speed and ability of search of algorithm, this paper adds local search (LS) and decomposition strategy (DS) to the traditional genetic algorithm (GA). The simulation shows that the proposed algorithm can provide more feasible non-dominated solutions and have faster computing speed.