The article directories

  • I. Theoretical basis
    • 1. Krill colony Algorithm (KH)
    • 2. Improved krill colony Algorithm (ANRKH)
      • (1) The time-varying nonlinear decreasing strategy of foraging weight and motion weight
      • (2) Random disturbance
      • (3) Natural selection
      • (4) ANRKH algorithm steps
  • 2. Simulation experiment and analysis
  • Iii. References
  • Four, Matlab simulation program

1. Krill colony Algorithm (KH)

Krill herd Algorithm (KH) is a new heuristic intelligent optimization algorithm, which is mainly based on the simulation study of the survival movement process of Antarctic Krill herd in the Marine environment. For each krill individual, its position update is mainly affected by three factors:

2. Improved krill colony Algorithm (ANRKH)

(1) The time-varying nonlinear decreasing strategy of foraging weight and motion weight

2. Simulation experiment and analysis

In order to verify the performance of ANRKH algorithm, ANRKH algorithm is compared with two related algorithms, KH algorithm and KHLD algorithm. Take F2, F3 and F5 in reference [1] for example. To ensure the fairness of the test algorithm, the common parameter Settings of the three algorithms are as follows: Number of individuals NP=100NP= 100NP=100, maximum number of iterations T max=1000 t_{Max}=1000tmax=1000, maximum induction speed Nmax=0.001 N^{Max}=0.001Nmax=0.001, Maximum foraging velocity Vf=0.02 V_f=0.02Vf=0.02, maximum random diffusion velocity D Max =0.005D ^{Max}=0.005Dmax=0.005; See Table 1 for different parameter values in the three algorithms. Each function was subjected to 30 independent numerical experiments.

Table 1 Parameter values in KH, KHLD, and ANRKH

The result is as follows:

Function: f2kh: maximum: 124.3161, minimum :28.6977, average :43.0134, standard deviation: 29.4007khld: maximum: 101.6416, min :26.2886, Mean :30.2278, SD :13.5046 ANRKH: Max: 81.6635, min :24.9959, mean :28.6219, SD :12.2347 Function: F3 KH: Max: 0.91102, Min :0.37651, Mean value :0.60851, SD: 0.12396khld: Max: 0.024535, Min :0.0063132, mean value :0.013014, SD :0.0046107 ANRKH: Max: 9.4045E-07, Min: 1.4627E-07, Mean: 5.5261E-07, standard deviation: 1.9059E-07 Function: F5KH: maximum: 0.27584, min :0.18773, Average :0.22487, standard deviation: 0.024735khld: Max: 0.038433, min :0.019181, average :0.027884, standard deviation :0.0054838 ANRKH: Max: 0.00044475, minimum value :0.00026492, average value :0.00033714, standard deviation: 4.5962E-05Copy the code

It can be seen that the ANRKH algorithm can effectively avoid falling into local optimization, and has significant advantages in global search and local exploration ability.

Iii. References

[1] Liu Pei, Gao Yuelin, GUO Wei. Improved krill colony algorithm based on Natural Selection and random perturbation [J]. Journal of Small Microcomputer Systems, 2017, 38(8): 1845-1849. [2] Li J , Tang Y , Hua C , et al. An improved krill herd algorithm: Krill herd with linear decreasing step[J]. Applied Mathematics & Computation, 2014, 234:356-367. [3] Gandomi A H, Alavi A H. Krill Herd:a new bio- inspired optimization algorithm[J]. Commun Nonlinear Sci Numer Simul, 2012, 17(12) : 4831-4845.