CN103679270A - Heuristic adaptive immune clonal method - Google Patents
Heuristic adaptive immune clonal method Download PDFInfo
- Publication number
- CN103679270A CN103679270A CN201310589584.6A CN201310589584A CN103679270A CN 103679270 A CN103679270 A CN 103679270A CN 201310589584 A CN201310589584 A CN 201310589584A CN 103679270 A CN103679270 A CN 103679270A
- Authority
- CN
- China
- Prior art keywords
- antibody
- affinity
- population
- tsc
- search
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Abstract
The invention discloses a heuristic adaptive immune clonal method. A novel clonal selection algorithm is provided by the aid of float-point encoding on the basis of a clonal selection theory. The heuristic adaptive immune clonal method has the advantages that an elite clonal variation operator and a heuristic crossover operator are defined; small-magnitude variation strategies are implemented on high-affinity antibodies, so that the high-affinity antibodies can be locally searched; heuristic crossover strategies are performed on intermediate-affinity antibody groups and high-affinity antibody groups to accelerate global search, and low-affinity antibodies die and are regenerated, so that the population diversity can be kept; affinity scale variation parameters can be adaptively adjusted in order to prevent evolution stagnation.
Description
The present invention is based on clonal selective theory, adopt floating-point encoding, propose a kind of new clonal selection algorithm.2 kinds of main operators of elite's clonal vaviation and heuristic intersection have been defined; High affinity antibody is implemented to small size Mutation Strategy to carry out Local Search, medium affinity antibody population is implemented to carry out the heuristic strategy intersecting to accelerate global search with high affinity antibody population, low affinity antibody is dead regenerates to keep population diversity; For preventing from evolving, stagnate, adjust adaptively affinity change of scale parameter.
For solving the problems of the technologies described above, technical scheme of the present invention is:
Immune Clonal Selection Algorithm is to form colony by clone, variation continuous passage, and the variation by antibody cloning group improves immunity (affinity, the adaptive value in genetic algorithm), thereby realizes the evolution of population.Although CLONALG has overcome the evolve problem of slow phenomenon of genetic algorithm local search ability poor and normal appearance, improves local search ability with less variation probability, has reduced ability of searching optimum; And although larger variation probability global optimizing ability improves, the precise decreasing of convergence.Therefore, its ability of searching optimum and local search ability are contradiction, find that a plurality of locally optimal solutions that CLONALG is used for searching for Solving Multimodal Function are effectively, but the precision of its globally optimal solution are lower and speed of convergence is slower in experiment.For this reason,
Based on clonal selective theory, adopt [0,1] floating-point encoding in interval, the search globally optimal solution of take is target, heuristic evolution thought under the guidance of elite's antibody is proposed, according to antibody, from the affinity of antigen, population is divided into the sub-population that function is different, the high affinity antibody fine search locally optimal solution that makes a variation in compared with small neighbourhood, medium affinity antibody carries out consistent the intersection (there is no variation) and accelerates overall coarse search with high affinity antibody, low affinity antibody is dead and random regenerates to keep population diversity; The degree of crossover and mutation is relevant with the affinity of antibody, and self-adaptation is adjusted affinity change of scale parameter in evolution, prevents from evolving and stagnates.
1 floating-point encoding
If g for antibody population A (g)=a1, a2,,, aN} is the N tuple of antibody aj, antibody aj=(zj1, zj2,,, zjl), i gene zji of j antibody is the floating number in [0,1] interval, zji and xji set up following mapping relations:
xji=pi+zji*(pi-qi)
Claim the floating-point encoding that zji is xji, the decoding that xji is zji.
2 populations are decomposed
A (g) according to the descending sort of antibody and antigen affinity size after, by 1B6B3 macro ratio, resolve into 3 sub-population A { m} (g), A{n} (g), A{r} (g), wherein A{m} (g) is the sub-population that affinity is the highest, the sub-population that A{n} (g) is medium affinity, A{r} (g) is the sub-population that affinity is minimum.(A(g)ISN@L,A(g)=A{m}(g)G?A{n}(g)GA{r}(g),m+n+r=N)。
3 clones (C loning)
Antagonist aIA{m} (g) GA{n} (g), clones by following formula:
C=GiCi=I@a,?i=1,2,…,nc
In formula: I is nc dimension row vector, and nc is clone's scale of antibody a, the clone body that Ci=a is a.
4 change of scale Tsc ()
If the sequence of antibody aIA (g) is rank (a), population scale is N, by formula (4), carries out affinity standardization: F (a)=(N-rank (a))/N
Obviously, the size of F (a) is relevant with target function value.Again F (a) is carried out to change of scale by formula (5).Tsc(a)=G#exp(-Q#F(a))
In formula: G and Q are constant, general 0<G<0.5,2[Q[10. can find out, and F (a) is larger, and Tsc (a) is less.
5 elite's clonal vaviation Tem ()
High affinity antibody (elite) aIA{m} (g) is positioned near majorized function f (xi) peak value, therefore should in less space, search for its locally optimal solution.Clone body CiIC to aIA{m} (g), carries out following consistent mutation operation by Probability p mI (0,1):
Cci=Ci+Tsc(a)*(2D-1)
In formula: D is the interval random number series vector of l dimension [0,1].Obviously, Ci makes a variation in Tsc (a) scope, and due to aIA{m} (g), Tsc (a) is less; In addition, pm gets the small value, general pm=1/l, and the variable that may morph is less, so its range of variation is also less.Like this, just limit elite's antibody and search for locally optimal solution in relatively little neighborhood, this locally optimal solution is also the optimum solution of current population.
6 heuristic clones are intersected THm ()
In order to improve rapidly affinity, the quickening antibody population search globally optimal solution of medium affinity antibody, in invention, adopt elite's antibody to carry out following heuristic intersection to medium affinity antibody.If the clone body CiIC of aIA{n} (g), chooses an antibody acIA{m} (g) at random, by formula (7), carry out interlace operation:
Cci=Ci@(1-Tsc(a)*D)+ac@(Tsc(a)*D)
In formula: D is the interval random number series vector of l dimension [0,1].Can find out, the antibody that affinity is larger is subject to the impact of elite's antibody less, and the less antibody of affinity is subject to the impact of elite's antibody larger.
The adaptive algorithm of 7 G values
G can think a whole enlargement factor, can find out, G gets large value, and Tsc (a) is worth corresponding increase, and the hunting zone of antibody is increased, and can accelerate global search, but local convergence is slower, the excessive random search that becomes; In like manner, the G local convergence precision that gets the small value is higher, but is easily absorbed in local minimum.For elite's antibody population A{m} (g), be mainly used in searching for locally optimal solution, therefore, G gets smaller value, and making it is G1; And antibody population A{n} (g) is mainly used in developing global optimum's solution space.Therefore, G gets higher value, and making it is G2.Generally there is G1<G2, be respectively used to the calculating of Tsc (a) value.
Compared with prior art, the invention has the beneficial effects as follows:
1) heuristic self-adaptation Immune Clonal Selection Algorithm is cloned principle based on artificial immunity, the elite who uses for reference genetic algorithm selects operator and crossover operator, propose elite's clonal vaviation operator, heuristic clone's crossover operator, made clonal selection algorithm accelerating global search.
2) can guarantee the precision of local convergence, its parameter adaptive mechanism can prevent Evolution of Population stagnation effectively.The test result of 4 complicated functions is shown to this algorithm has overcome premature convergence problem effectively, fast convergence rate, stable performance, precision is high.
3) simulation comparison test shows that this algorithm has obviously improved CLONALG algorithm, and calculated amount is little, stable performance, and precision is high, and being difficult for being absorbed in the Solve problems that this algorithm of local optimum is complicated continuous function provides an effective way.
Claims (1)
1. a heuristic self-adaptation immune clone method is divided into following components: Immune Clonal Selection Algorithm is to form colony by clone, variation continuous passage, variation by antibody cloning group improves immunity (affinity, adaptive value in genetic algorithm), thus realize the evolution of population; Although CLONALG has overcome the evolve problem of slow phenomenon of genetic algorithm local search ability poor and normal appearance, improves local search ability with less variation probability, has reduced ability of searching optimum; And although larger variation probability global optimizing ability improves, the precise decreasing of convergence; Therefore, its ability of searching optimum and local search ability are contradiction, find that a plurality of locally optimal solutions that CLONALG is used for searching for Solving Multimodal Function are effectively, but the precision of its globally optimal solution are lower and speed of convergence is slower in experiment; For this reason, based on clonal selective theory, adopt [0,1] floating-point encoding in interval, the search globally optimal solution of take is target, heuristic evolution thought under the guidance of elite's antibody is proposed, according to antibody, from the affinity of antigen, population is divided into the sub-population that function is different, the high affinity antibody fine search locally optimal solution that makes a variation in compared with small neighbourhood, medium affinity antibody carries out consistent the intersection (there is no variation) and accelerates overall coarse search with high affinity antibody, low affinity antibody is dead and random regenerates to keep population diversity; The degree of crossover and mutation is relevant with the affinity of antibody, and self-adaptation is adjusted affinity change of scale parameter in evolution, prevents from evolving and stagnates;
Floating-point encoding
If g for antibody population A (g)=a1, a2,,, aN} is the N tuple of antibody aj, antibody aj=(zj1, zj2,,, zjl), i gene zji of j antibody is the floating number in [0,1] interval, zji and xji set up following mapping relations:
xji=pi+zji*(pi-qi)
Claim the floating-point encoding that zji is xji, the decoding that xji is zji;
Population is decomposed
A (g) according to the descending sort of antibody and antigen affinity size after, by 1B6B3 macro ratio, resolve into 3 sub-population A { m} (g), A{n} (g), A{r} (g), wherein A{m} (g) is the sub-population that affinity is the highest, the sub-population that A{n} (g) is medium affinity, A{r} (g) is the sub-population that affinity is minimum; (A (g) ISN@L, A (g)=A{m} (g) G A{n} (g) GA{r} (g), m+n+r=N);
Clone (C loning)
Antagonist aIA{m} (g) GA{n} (g), clones by following formula:
C=GiCi=I@a,?i=1,2,…,nc
In formula: I is nc dimension row vector, and nc is clone's scale of antibody a, the clone body that Ci=a is a;
Change of scale Tsc ()
If the sequence of antibody aIA (g) is rank (a), population scale is N, by formula (4), carries out affinity standardization: F (a)=(N-rank (a))/N
Obviously, the size of F (a) is relevant with target function value; Again F (a) is carried out to change of scale by formula (5); Tsc (a)=G#exp (Q#F (a))
In formula: G and Q are constant, general 0<G<0.5,2[Q[10. can find out, and F (a) is larger, and Tsc (a) is less;
Elite's clonal vaviation Tem ()
High affinity antibody (elite) aIA{m} (g) is positioned near majorized function f (xi) peak value, therefore should in less space, search for its locally optimal solution; Clone body CiIC to aIA{m} (g), carries out following consistent mutation operation by Probability p mI (0,1):
Cci=Ci+Tsc(a)*(2D-1)
In formula: D is the interval random number series vector of l dimension [0,1]; Obviously, Ci makes a variation in Tsc (a) scope, and due to aIA{m} (g), Tsc (a) is less; In addition, pm gets the small value, general pm=1/l, and the variable that may morph is less, so its range of variation is also less; Like this, just limit elite's antibody and search for locally optimal solution in relatively little neighborhood, this locally optimal solution is also the optimum solution of current population;
Heuristic clone is intersected THm ()
In order to improve rapidly affinity, the quickening antibody population search globally optimal solution of medium affinity antibody, in invention, adopt elite's antibody to carry out following heuristic intersection to medium affinity antibody; If the clone body CiIC of aIA{n} (g), chooses an antibody acIA{m} (g) at random, by formula (7), carry out interlace operation:
Cci=Ci@(1-Tsc(a)*D)+ac@(Tsc(a)*D)
In formula: D is the interval random number series vector of l dimension [0,1]; Can find out, the antibody that affinity is larger is subject to the impact of elite's antibody less, and the less antibody of affinity is subject to the impact of elite's antibody larger;
The adaptive algorithm of G value
G can think a whole enlargement factor, can find out, G gets large value, and Tsc (a) is worth corresponding increase, and the hunting zone of antibody is increased, and can accelerate global search, but local convergence is slower, the excessive random search that becomes; In like manner, the G local convergence precision that gets the small value is higher, but is easily absorbed in local minimum; For elite's antibody population A{m} (g), be mainly used in searching for locally optimal solution, therefore, G gets smaller value, and making it is G1; And antibody population A{n} (g) is mainly used in developing global optimum's solution space; Therefore, G gets higher value, and making it is G2; Generally there is G1<G2, be respectively used to the calculating of Tsc (a) value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310589584.6A CN103679270A (en) | 2013-11-21 | 2013-11-21 | Heuristic adaptive immune clonal method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310589584.6A CN103679270A (en) | 2013-11-21 | 2013-11-21 | Heuristic adaptive immune clonal method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103679270A true CN103679270A (en) | 2014-03-26 |
Family
ID=50316754
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310589584.6A Pending CN103679270A (en) | 2013-11-21 | 2013-11-21 | Heuristic adaptive immune clonal method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103679270A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104281972A (en) * | 2014-10-23 | 2015-01-14 | 成都信息工程学院 | Social network false information control method based on clonal selection algorithm |
| CN111382896A (en) * | 2018-12-29 | 2020-07-07 | 陕西师范大学 | WTA target optimization method of adaptive chaotic parallel clonal selection algorithm |
| CN113965358A (en) * | 2021-09-28 | 2022-01-21 | 石河子大学 | Network security detection method and system for comprehensive energy system |
-
2013
- 2013-11-21 CN CN201310589584.6A patent/CN103679270A/en active Pending
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104281972A (en) * | 2014-10-23 | 2015-01-14 | 成都信息工程学院 | Social network false information control method based on clonal selection algorithm |
| CN104281972B (en) * | 2014-10-23 | 2017-06-09 | 成都信息工程学院 | Social networks deceptive information control method based on clonal selection algorithm |
| CN111382896A (en) * | 2018-12-29 | 2020-07-07 | 陕西师范大学 | WTA target optimization method of adaptive chaotic parallel clonal selection algorithm |
| CN111382896B (en) * | 2018-12-29 | 2023-10-31 | 陕西师范大学 | WTA target optimization method of self-adaptive chaotic parallel clone selection algorithm |
| CN113965358A (en) * | 2021-09-28 | 2022-01-21 | 石河子大学 | Network security detection method and system for comprehensive energy system |
| CN113965358B (en) * | 2021-09-28 | 2023-04-28 | 石河子大学 | Network security detection method and system for comprehensive energy system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103324983B (en) | A kind of isomorphism identification method of mechanism kinematics chain based on immune genetic hybrid algorithm | |
| CN107609683B (en) | Firefly algorithm-based cascade reservoir group scheduling optimization method | |
| CN110598983A (en) | Cascade reservoir optimal scheduling method of self-adaptive improved particle swarm algorithm | |
| Singh et al. | Performance of a hybrid EA-DE-memetic algorithm on CEC 2011 real world optimization problems | |
| CN103679270A (en) | Heuristic adaptive immune clonal method | |
| CN106777449A (en) | Distribution Network Reconfiguration based on binary particle swarm algorithm | |
| CN105976048A (en) | Power transmission network extension planning method based on improved artificial bee colony algorithm | |
| CN110119408A (en) | Mobile object continuous-query method under geographical space real-time streaming data | |
| CN101616074A (en) | Multicast routing optimization method based on quantum evolution | |
| Chen et al. | Application of novel clonal algorithm in multiobjective optimization | |
| CN110635504A (en) | Power distribution network distributed photovoltaic absorption capacity evaluation method based on improved pollen algorithm | |
| Wang et al. | ACSAMO: an adaptive multiobjective optimization algorithm using the clonal selection principle | |
| Pattanaik et al. | Opposition-based differential evolution for hydrothermal power system | |
| CN105678623A (en) | Metaheuristic searching method for solving flexibility workshop job scheduling | |
| CN106611217A (en) | Weighted genetic local search algorithm for multi-objective flow shop scheduling | |
| CN112183843A (en) | Thermal power plant load optimal distribution method based on hybrid intelligent algorithm | |
| Yuan et al. | Particle swarm optimisation algorithm with forgetting character | |
| CN106647757B (en) | Based on the robot path planning method for combining individual differential evolution | |
| Fu et al. | A hybrid artificial immune network with swarm learning | |
| CN103646285A (en) | Chaos-clonal evolution method | |
| Fang et al. | Multi-objective optimal allocation for regional water and land resources | |
| Zhao et al. | Hybrid algorithm for solving traveling salesman problem | |
| Mirsadeghi et al. | Hybridizing artificial bee colony with simulated annealing | |
| CN108921277A (en) | A kind of multi-modal optimization backbone particle swarm optimization based on improvement species conversation strategy | |
| Liu et al. | Multi-objective reactive power optimization by Modified Artificial Fish Swarm Algorithm in IEEE 57-bus power system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20140326 |
|
| WD01 | Invention patent application deemed withdrawn after publication |