In this paper a comparison of single and multi-population evolutionary algorithm is presented. Tested algorithms are used to determine close to optimal ship paths in collision avoidance situation. For this purpose a p...
详细信息
ISBN:
(纸本)9783037859209
In this paper a comparison of single and multi-population evolutionary algorithm is presented. Tested algorithms are used to determine close to optimal ship paths in collision avoidance situation. For this purpose a path planning problem is defined. A specific structure of the individual path and fitness function is presented. Principle of operation of single-population and multi-population evolutionary algorithm is described. Using presented algorithms the simulations on three close to real sea environments were performed. Regardless of the test situation constant time simulation was maintained. Obtained results are presented in graphical form (sequences of successive stages of the simulation) and in form of table in which the values of fitness function for best individual in each simulation were compared. Undertaken research allow to select evolutionary algorithm that, assuming constant simulation time, will determine a better path in close to real collision avoidance situation at sea.
Shuffled multi-population Bat algorithm (SMPBat) is a recently proposed hybrid variant of bat algorithm. It incorporates the strengths of two recent variants of bat algorithm-Enhanced Shuffled Bat Algorithm and Bat al...
详细信息
ISBN:
(纸本)9781538653142
Shuffled multi-population Bat algorithm (SMPBat) is a recently proposed hybrid variant of bat algorithm. It incorporates the strengths of two recent variants of bat algorithm-Enhanced Shuffled Bat Algorithm and Bat algorithm with Ring Master-Slave strategy. SMPBat hybridizes the sub-population generation and manipulation mechanism of the two algorithms to device an enhanced variant of BA. There are multiple parameters controlling the flow of execution of SMPBat. These parameters are set at the beginning of the execution of the algorithm. This paper proposes incorporation of multiparameter setting into SMPBat, where different sub-populations work with different sets of parameter values. Additionally, the impact of a refined search mechanism is also studied. The proposed variants are tested over 20 benchmark functions and a real-world optimization problem. Results establish the robustness of the proposed work.
multi -population -based methods are widely employed for solving constrained multiobjective optimization problems (CMOPs). The population collaboration strategy is a critical part of multi -populationalgorithms, and ...
详细信息
multi -population -based methods are widely employed for solving constrained multiobjective optimization problems (CMOPs). The population collaboration strategy is a critical part of multi -populationalgorithms, and different collaboration strategies perform well on different complex CMOPs. However, these single -population collaboration strategies are still challenging to adapt to various CMOPs with different characteristics. To address this issue, we propose a novel tri-population hybrid collaboration evolutionary algorithm called TPHCEA, which includes a constraint -relaxed population (denoted as mainpop ), a constraint -ignored auxiliary population (denoted as auxpop 1 ), and an auxiliary population (denoted as auxpop 2 ) for the original CMOP, to search optimal solutions in the feasible region. Specifically, due to the different complementarities of the two auxiliary populations, mainpop collaborates with auxpop 1 and auxpop 2 in a dynamic choice between strong and weak cooperation. The effectiveness of TP-HCEA is validated through comparative analysis with seven state-of-the-art algorithms in four CMOP benchmark suites and nine real -world problems.
Competition and cooperation are powerful metaphors that have informed improvements in multi-population algorithms such as the Cooperative Coevolutionary Genetic Algorithm, Cooperative Particle Swarm Optimization, and ...
详细信息
ISBN:
(纸本)9781450356183
Competition and cooperation are powerful metaphors that have informed improvements in multi-population algorithms such as the Cooperative Coevolutionary Genetic Algorithm, Cooperative Particle Swarm Optimization, and Factored Evolutionary algorithms (FEA). However, we suggest a different perspective can give a finer grained understanding of how multi-population algorithms come together to avoid problems like hitchhiking and pseudo-minima. In this paper, we apply the concepts of information sharing and conflict resolution through Pareto improvements to analyze the distributed version of FEA (DFEA). As a result, we find the original DFEA failed to implement FEA with complete fidelity. We then revise DFEA and examine the differences between it and FEA and the new implications for relaxing consensus in the distributed algorithm.
We propose an evolutionary algorithm with a novel diversity oscillation mechanism for the Capacitated Vehicle Routing Problem with Time Windows (CVRPTW). Evolutionary algorithms are among state-of-the-art methods for ...
详细信息
ISBN:
(纸本)9783031147142;9783031147135
We propose an evolutionary algorithm with a novel diversity oscillation mechanism for the Capacitated Vehicle Routing Problem with Time Windows (CVRPTW). Evolutionary algorithms are among state-of-the-art methods for vehicle routing problems and the diversity management is the key component of many of these algorithms. In our algorithm the diversity level slowly oscillates between its minimum and maximum value, however, whenever a new best solution is found the algorithm switches to decreasing the diversity level in order to intensify the search in the vicinity of the new best solution. We use also an additional population of high quality diverse solutions, which may be used to re-fill the main population when the diversification level is increased. The results of the computational experiment indicate that the proposed mechanism significantly improves the performance of our hybrid evolutionary algorithm on typical CVRPTW benchmarks and that the proposed algorithm is competitive to the state-of-the-art results presented in the literature.
Flower Pollination Algorithm (FPA) is a bio-inspired metaheuristic that simulates pollination behavior of flowers. FPA is introduced to solve global optimization problems. Subsequently, it has been applied to a variet...
详细信息
Flower Pollination Algorithm (FPA) is a bio-inspired metaheuristic that simulates pollination behavior of flowers. FPA is introduced to solve global optimization problems. Subsequently, it has been applied to a variety of problems. The present study introduces some new extensions and modifications for FPA. In this respect, first, abiotic pollination mechanism of FPA is modified. Secondarily, in order to control convergence speed, a step size function that is used in both global and local pollination along with the randomness factor is adopted. Finally, FPA is extended as a species-based algorithm by partitioning whole population into smaller-sized groups that independently search for promising regions. Performances of the proposed extensions are analyzed by using the well-known unconstrained function optimization problems and Morrison and De Jong's field of cones function. Finally, non parametric statistical tests are conducted to demonstrate possible significant improvements over standard FPA. As shown by these statistically verified results, the first FPA modification with the proposed selection mechanism and step size function achieves the best results in global optimization problems while the species-based FPA modification is found as a promising algorithm to solve multi-modal problems of De Jong's field of cones function. (C) 2021 Elsevier B.V. All rights reserved.
暂无评论