基于克隆选择算法的非线性优化迭代学习控制；Clonal selection algorithm based nonlinear optimal iterative learning control

为实现对模型不确定的有约束非线性系统在特定时间域上输出轨迹的有效跟踪,将改进的克隆选择算法用于求解迭代学习控制中的优化问题。提出基于克隆选择算法的非线性优化迭代学习控制。在每次迭代运算后,一个克隆选择算法用于求解下次迭代运算中的最优输入,另一个克隆选择算法用于修正系统参考模型。仿真结果表明,该方法比GA-ILC具有更快的收敛速度,能够有效处理输入上的约束以及模型不确定问题,通过少数几次迭代学习就能取得满意的跟踪效果。

On Clonal Selection.

Clonal selection has been a dominant theme in many immune-inspired algorithms applied to machine learning and optimisation. We examine existing clonal selections algorithms for learning from a theoertical and empirical perspective and assert that the widely accepted computational interpretation of clonal selection is compromised both algorithmically andbiologically. We suggest a more capable abstraction of the clonal selection principle grounded in probabilistic estimation and approximation and demonstrate how it addresses some of the shortcomings in existing algorithms. We further show that by recasting black-box optimisation as a learning problem, the same abstraction may be re-employed; thereby taking steps toward unifying the clonal selection principle and distinguishing it from natural selection.

From clonal selection to immune networks: induction of silent idiotypes.

Journal Article

Differential evolution combined with clonal selection for dynamic economic dispatch

Dynamic economic load dispatch (DELD) is one of the most important steps in power system operation. Various optimisation algorithms for solving the problem have been developed; however, due to the non-convex characteristics and large dimensionality of the problem, it is necessary to explore new methods to further improve the dispatch results and minimise the costs. This article proposes a hybrid differential evolution (DE) algorithm, namely clonal selection-based differential evolution (CSDE), to solve the problem. CSDE is an artificial intelligence technique that can be applied to complex optimisation problems which are for example nonlinear, large scale, non-convex and discontinuous. This hybrid algorithm combines the clonal selection algorithm (CSA) as the local search technique to update the best individual in the population, which enhances the diversity of the solutions and prevents premature convergence in DE. Furthermore, we investigate four mutation operations which are used in CSA as the hyper-mutation operations. Finally, an efficient solution repair method is designed for DELD to satisfy the complicated equality and inequality constraints of the power system to guarantee the feasibility of the solutions. Two benchmark power systems are used to evaluate the performance of the proposed method. The experimental results show that the proposed CSDE/best/1 approach significantly outperforms nine other variants of CSDE and DE, as well as most other published methods, in terms of the quality of the solution and the convergence characteristics.

Optimal planning of cocoa clonal selection programmes

Data from three cocoa (Theobroma cacao) clonal selection trials are used to investigate the genetic and environmental components of variation in yield and the percentage of total pods affected by black pod disease (Phytophtora pod rot). Simulations based on these estimated components of variation are then used to discuss the best choice in future of numbers of clones, replicates and years of harvest to maximise selection advances in the traits measured. The three main conclusions are the need to increase the number of clones at the expense of the number of replicates of each clone, the diminishing returns from additional years of harvesting and the importance of widening the genetic base of the clones chosen to be tested.

Spectral element model updating for damage identification using clonal selection algorithm

A spectral element model updating procedure is presented to identify damage in a structure using Guided wave propagation results. Two damage spectral elements (DSE1 and DSE2) are developed to model the local (cracks in reinforcement bar) and global (debonding between reinforcement bar and concrete) damage in one-dimensional homogeneous and composite waveguide, respectively. Transfer matrix method is adopted to assemble the stiffness matrix of multiple spectral elements. In order to solve the inverse problem, clonal selection algorithm is used for the optimization calculations. Two displacement-based functions and two frequency-based functions are used as objective functions in this study. Numerical simulations of wave propagation in a bare steel bar and in a reinforcement bar without and with various assumed damage scenarios are carried out. Numerically simulated data are then used to identify local and global damage of the steel rebar and the concrete-steel interface using the proposed method. Results show that local damage is easy to be identified by using any considered objective function with the proposed method while only using the wavelet energy-based objective function gives reliable identification of global damage. The method is then extended to identify multiple damages in a structure. To further verify the proposed method, experiments of wave propagation in a rectangular steel bar before and after damage are conducted. The proposed method is used to update the structural model for damage identification. The results demonstrate the capability of the proposed method in identifying cracks in steel bars based on measured wave propagation data.

Roles of the Immune Cytokine Environment on Clonal Growth of Murine and Human Tet2 Mutant Bone Marrow Progenitors: Implications for Myelodysplastic Syndromes

TET2 is a tumor suppressor gene that has been implicated in the epigenetic regulation of gene expression. Inactivating TET2 mutations are common in MDS. These mutations may contribute to early clonal dominance and myeloid transformation, although the exact mechanisms remain to be elucidated. Common to the environment of MDS are elevations in cytokines, such as TNFα and IFN-γ. It was hypothesized that inflammatory cytokines TNF-α and IFN-γ may promote clonal expansion of TET2 mutant progenitors. Adult (10-14 weeks-old) Tet2 wild type (+/+) and Tet2 mutant (-/-) C57BL/6 mice strains were chosen as a model system. Lineage negative cells (Lin-), enriched for hematopoietic stem and progenitor cells, were isolated from Tet2 +/+ and -/- bone marrow and cultured in the absence or presence of varying concentrations of TNFα or IFN-γ in methylcellulose colony formation assays and long term cell culture assays, over a period of 12 and 30 days respectively, and their colony growth, cell count, immunophenotype and resistance to apoptosis were examined. Where indicated, serial re-plating was performed. Expression of apoptotic regulators was assessed by qRT-PCR. In the triplicate experiments, starting with equal densities of Tet2 +/+ and -/- Lin- cells, Tet2 -/- Lin- cells displayed increased resistance to cytokine-induced growth suppression and superior colony forming ability over +/+ in the serial re-plating assays under stress of increasing TNFα or IFN γ. Tet2 -/- progenitors also displayed a lower apoptotic index compared to +/+ under stress of increasing TNFα, suggesting increased resistance to TNFα induced apoptosis. Transcriptional data showed low expression of Tnfr1, Fas and caspase 8, as well as a high expression of Bcl-2 and Iap1 in Tet2 -/- compared to +/+ under stress of TNFα. Tet2-/- also showed increased basal expression of endogenous TNFα mRNA compared to +/+. In the human colony growth assay, the clonal growth of TET2 mutant CFU-GM progenitors was enhanced at low TNFα concentrations. Conclusion: Mutations that promote resistance to environmental stem cell stressors are a known mechanism of clonal selection in aplastic anaemia and JAK2-mutant MPN and our findings suggest that this mechanism may be critical to clonal selection and dominance in MDS.

A human-simulated immune evolutionary computation approach

Premature convergence to local optimal solutions is one of the main difficulties when using evolutionary algorithms in real-world optimization problems. To prevent premature convergence and degeneration phenomenon, this paper proposes a new optimization computation approach, human-simulated immune evolutionary algorithm (HSIEA). Considering that the premature convergence problem is due to the lack of diversity in the population, the HSIEA employs the clonal selection principle of artificial immune system theory to preserve the diversity of solutions for the search process. Mathematical descriptions and procedures of the HSIEA are given, and four new evolutionary operators are formulated which are clone, variation, recombination, and selection. Two benchmark optimization functions are investigated to demonstrate the effectiveness of the proposed HSIEA.

Hormone resistance, invasiveness, and metastatic potential in breast cancer

Critical phenotypic changes that occur during the progression of breast cancer include the loss of hormone-dependence, acquired resistance to systemic therapies, and increased metastatic potential. We have isolated a series of MCF-7 human breast cancer variants which exhibit hormone-independent growth, antiestrogen resistance, and increased metastatic potential. Analysis of the phenotypes of these variants strongly suggests that changes in the expression of specific genes may be critical to the generation of phenotypic diversity in the process of malignant progression in breast cancer. Epigenetic changes may contribute significantly to the generation of these phenotypic changes observed during breast cancer progression. Many of the characteristics of the progressed phenotypes appear to have arisen in response to appropriate selective pressures (growth in ovariectomized nude mice; growth in the presence of antiestrogens). These observations are consistent with the concept of clonal selection and expansion in the process of malignant progression.

Relationships between Indole-3-butyric Acid, Photoinhibition and Adventitious rooting of Corymbia torelliana, C. citriodora and F1 Hybrid Cuttings

Rooted cutting propagation is widely used for maximising tree yield, quality and uniformity in conjunction with clonal selection. Some eucalypt species are deployed as rooted cuttings but many are considered to difficult to root. This study examined IBA effects on photoinhibition, root formation, mortality and root and shoot development in cuttings of Corymbia torelliana, C. citriodora and their hybrids. IBA had little or no effect on photoinhibition but it had strong, dose-dependent effects on root formation and mortality. IBA frequently increases primary root number of rooted cutting but it did not increase total root weight, length, surface area or volume, possibly because the highest doe (8g IBA/kg IBA/kg powder) caused leaf abscission and sometimes reduced leaf area (by 55-79%)or shoot dry weight (by 40-58%). An intermediate dose (3g IBA/kg powder) most consistnely improved root formation with little or no effect on mortality or shoot development. Across the F1 hybrid families this treatment increased the number of rooted cuttings by 72-121% and more than ddoubled the number of primary roots per rooted cutting (from 1.1-1.7 roots to 3.5-4.1 roots). This simple treatment will facilitate commercial multiplication of superior individuals or selected families of C. torelliana x C. citriodora through a vegetative propagation system.

An Optimal Fuzzy Logic Controller tuned with Artificial Immune System

In this paper, a method for the tuning the membership functions of a Mamdani type Fuzzy Logic Controller (FLC) using the Clonal Selection Algorithm(CSA) a model of the Artificial Immune System(AIS) paradigm is examined. FLC's are designed for two problems, firstly the linear cart centering problem and secondly the highly nonlinear inverted pendulum problem. The FLC tuned by AIS is compared with FLC tuned by GA. In order to check the robustness of the designed PLC's white noise was added to the system, further, the masses of the cart and the length and mass of the pendulum are changed. The PLC's were also tested in the presence of faulty rules. Finally, Kruskal Wallis test was performed to compare the performance of the GA and AIS. An insight into the algorithms are also given by studying the effect of the important parameters of GA and AIS.

How do we evaluate Artificial Immune Systems?

RAE2008

Specificity enhancement of polyclonal antisera by the induction of tolerance to unwanted cross-reacting determinants

Steroids form a structurally closely related group. As a result, antibodies produced for use in immunoassays regularly show unwanted cross-reactivities, These may be reduced by altering hapten-protein coupling procedures, thereby reducing the exposure of the determinants giving rise to the undesirable cross-reaction. However, these procedures carl prove to be complex, expensive and nor totally predictable in outcome. Exploitation of the clonal selection theory is an attractive alternative approach. The host is primed with the interfering cross-reactant coupled to a non-immunogenic amino acid copolymer to inactivate the B-lymphocyte clones specific for this steroid, producing a specific immunotolerance. Then, 3 days Inter, the host is immunized with the steroid against which nn antibody is required. The clones producing antibody to this immunogen are unaffected and the cross-reactivity is significantly reduced or deleted The technique has been applied to the reduction of endogenous sex steroid cross-reactivity from antibodies prepared against synthetic and semi-synthetic androgens (17 alpha-methyltestosterone, 19-nor-beta-testosterone) and the progestogen medroxyprogesterone. Antibodies prepared against the synthetic oestrogen zeranol using this technique have significantly reduced its undesirable cross-reactivity with the fungal metabolite 7 alpha-zearalenol. Highly specific antisera have been generated in all cases, the only adverse effect being a reduction in the titres achieved in comparison with rabbits receiving the conventional immunizing regime.

A robust controller design method for feedback substitution schemes using genetic algorithms

Controllers for feedback substitution schemes demonstrate a trade-off between noise power gain and normalized response time. Using as an example the design of a controller for a radiometric transduction process subjected to arbitrary noise power gain and robustness constraints, a Pareto-front of optimal controller solutions fulfilling a range of time-domain design objectives can be derived. In this work, we consider designs using a loop shaping design procedure (LSDP). The approach uses linear matrix inequalities to specify a range of objectives and a genetic algorithm (GA) to perform a multi-objective optimization for the controller weights (MOGA). A clonal selection algorithm is used to further provide a directed search of the GA towards the Pareto front. We demonstrate that with the proposed methodology, it is possible to design higher order controllers with superior performance in terms of response time, noise power gain and robustness.