An Estimation of Distribution Algorithm for Nurse Scheduling

Schedules can be built in a similar way to a human scheduler by using a set of rules that involve domain knowledge. This paper presents an Estimation of Distribution Algorithm (EDA) for the nurse scheduling problem, which involves choosing a suitable scheduling rule from a set for the assignment of each nurse. Unlike previous work that used Genetic Algorithms (GAs) to implement implicit learning, the learning in the proposed algorithm is explicit, i.e. we identify and mix building blocks directly. The EDA is applied to implement such explicit learning by building a Bayesian network of the joint distribution of solutions. The conditional probability of each variable in the network is computed according to an initial set of promising solutions. Subsequently, each new instance for each variable is generated by using the corresponding conditional probabilities, until all variables have been generated, i.e. in our case, a new rule string has been obtained. Another set of rule strings will be generated in this way, some of which will replace previous strings based on fitness selection. If stopping conditions are not met, the conditional probabilities for all nodes in the Bayesian network are updated again using the current set of promising rule strings. Computational results from 52 real data instances demonstrate the success of this approach. It is also suggested that the learning mechanism in the proposed approach might be suitable for other scheduling problems.

A Pyramidal Evolutionary Algorithm with Different Inter-Agent Partnering Strategies for Scheduling Problems

This paper combines the idea of a hierarchical distributed genetic algorithm with different inter-agent partnering strategies. Cascading clusters of sub-populations are built from bottom up, with higher-level sub-populations optimising larger parts of the problem. Hence higher-level sub-populations search a larger search space with a lower resolution whilst lower-level sub-populations search a smaller search space with a higher resolution. The effects of different partner selection schemes amongst the agents on solution quality are examined for two multiple-choice optimisation problems. It is shown that partnering strategies that exploit problem-specific knowledge are superior and can counter inappropriate (sub-) fitness measurements.

An Estimation of Distribution Algorithm for Nurse Scheduling

Schedules can be built in a similar way to a human scheduler by using a set of rules that involve domain knowledge. This paper presents an Estimation of Distribution Algorithm (EDA) for the nurse scheduling problem, which involves choosing a suitable scheduling rule from a set for the assignment of each nurse. Unlike previous work that used Genetic Algorithms (GAs) to implement implicit learning, the learning in the proposed algorithm is explicit, i.e. we identify and mix building blocks directly. The EDA is applied to implement such explicit learning by building a Bayesian network of the joint distribution of solutions. The conditional probability of each variable in the network is computed according to an initial set of promising solutions. Subsequently, each new instance for each variable is generated by using the corresponding conditional probabilities, until all variables have been generated, i.e. in our case, a new rule string has been obtained. Another set of rule strings will be generated in this way, some of which will replace previous strings based on fitness selection. If stopping conditions are not met, the conditional probabilities for all nodes in the Bayesian network are updated again using the current set of promising rule strings. Computational results from 52 real data instances demonstrate the success of this approach. It is also suggested that the learning mechanism in the proposed approach might be suitable for other scheduling problems.

Articulation and Clarification of the Dendric Cell Algorithm

The Dendritic Cell algorithm (DCA) is inspired by recent work in innate immunity. In this paper a formal description of the DCA is given. The DCA is described in detail, and its use as an anomaly detector is illustrated within the context of computer security. A port scan detection task is performed to substantiate the influence of signal selection on the behaviour of the algorithm. Experimental results provide a comparison of differing input signal mappings.

Influence of phosphine resistance genes on flight propensity and resource location in Tribolium castaneum (Herbst) (Coleoptera: Tenebrionidae): the landscape for selection

Phosphine resistance in Tribolium castaneum Herbst (Coleoptera: Tenebrionidae) has evolved through changes to enzymes involved in basic metabolic pathways. These changes impose metabolic stress and could affect energy-demanding behaviours. We therefore tested whether phosphine resistance alleles impact the movement of these insects in their quest for new resources. We measured walking and flight parameters of four T. castaneum genotypes: (1) a field-derived population, (2) a laboratory cultured, phosphine-susceptible reference strain, (3) a laboratory cultured, phosphine-resistant reference strain, and (4) a resistant introgressed strain that is almost identical genetically to the susceptible population. The temporal pattern of flight was identical across all populations, but resistant beetles took flight significantly less, walked more slowly, and located resources less successfully than did susceptible beetles. Also, the field-derived beetles (proved not to be carrying resistance genes) walked significantly faster and more directly towards food resources, and had a higher propensity for flight when compared to the susceptible laboratory beetles. These negative effects suggest survival of beetles with the resistance alleles will be compromised should they leave phosphine application sites. The field for selection therefore extends beyond the site at which phosphine fumigant imposed its effect, and other mutations are also likely to be affected in this way.

Rootstock selection, nitrogen and calcium influence postharvest disease in avocado

Field trials evaluating several parameters of growth, fruit yield and quality of 'Hass' avocado grafted to different rootstocks were established in 2004-2005 in four different growing regions of Australia. Fruit were harvested in three seasons from 2008, ripened and assessed for severity and incidence of anthracnose and stem end rot diseases. Peel samples were collected at harvest and analysed for concentrations of the cations (N, K, Ca, Mg). Rootstock significantly affected marketability of fruit (no stem end rot and less than 5% anthracnose) in 58% of the total number of trials evaluated, with better quality fruit harvested from 'Hass' grafted to Guatemalan or West Indian rootstocks such as 'A10' or 'Velvick'. Fruit quality was frequently poor from trees grafted to Mexican race rootstocks, regardless of growing location. Correlation analyses showed that fruit from rootstocks with superior fruit quality was often associated with lower skin N and higher Ca concentrations. There were significant positive correlations between anthracnose and skin N or N:Ca ratio in 75% of trials evaluated. There was a significant negative correlation between anthracnose and Ca in 42% of trials. The correlations between stem end rot and skin N (positive) or Ca (negative) were each significant in 42% of trials. Based on the results in this project, N:Ca ratios in the skin of unripe avocado fruit at harvest may provide one of the best indicators of potential postharvest disease in ripe fruit, and may have implications for fertiliser regimes.

'A Bayesian Optimisation Algorithm for the Nurse Scheduling Problem'

Abstract- A Bayesian optimization algorithm for the nurse scheduling problem is presented, which involves choosing a suitable scheduling rule from a set for each nurse's assignment. Unlike our previous work that used GAs to implement implicit learning, the learning in the proposed algorithm is explicit, i.e. eventually, we will be able to identify and mix building blocks directly. The Bayesian optimization algorithm is applied to implement such explicit learning by building a Bayesian network of the joint distribution of solutions. The conditional probability of each variable in the network is computed according to an initial set of promising solutions. Subsequently, each new instance for each variable is generated by using the corresponding conditional probabilities, until all variables have been generated, i.e. in our case, a new rule string has been obtained. Another set of rule strings will be generated in this way, some of which will replace previous strings based on fitness selection. If stopping conditions are not met, the conditional probabilities for all nodes in the Bayesian network are updated again using the current set of promising rule strings. Computational results from 52 real data instances demonstrate the success of this approach. It is also suggested that the learning mechanism in the proposed approach might be suitable for other scheduling problems.

Partnering Strategies for Fitness Evaluation in a Pyramidal Evolutionary Algorithm

This paper combines the idea of a hierarchical distributed genetic algorithm with different inter-agent partnering strategies. Cascading clusters of sub-populations are built from bottom up, with higher-level sub-populations optimising larger parts of the problem. Hence higher-level sub-populations search a larger search space with a lower resolution whilst lower-level sub-populations search a smaller search space with a higher resolution. The effects of different partner selection schemes for (sub-)fitness evaluation purposes are examined for two multiple-choice optimisation problems. It is shown that random partnering strategies perform best by providing better sampling and more diversity.

A Pyramidal Evolutionary Algorithm with Different Inter-Agent Partnering Strategies for Scheduling Problems

This paper combines the idea of a hierarchical distributed genetic algorithm with different inter-agent partnering strategies. Cascading clusters of sub-populations are built from bottom up, with higher-level sub-populations optimising larger parts of the problem. Hence higher-level sub-populations search a larger search space with a lower resolution whilst lower-level sub-populations search a smaller search space with a higher resolution. The effects of different partner selection schemes amongst the agents on solution quality are examined for two multiple-choice optimisation problems. It is shown that partnering strategies that exploit problem-specific knowledge are superior and can counter inappropriate (sub-) fitness measurements.

A Pyramidal Genetic Algorithm for Multiple-Choice Problems

This paper combines the idea of a hierarchical distributed genetic algorithm with different inter-agent partnering strategies. Cascading clusters of sub-populations are built from bottom up, with higher-level sub-populations optimising larger parts of the problem. Hence, higher-level sub-populations search a larger search space with a lower resolution whilst lower-level sub-populations search a smaller search space with a higher resolution. The effects of different partner selection schemes amongst the agents on solution quality are examined for two multiple-choice optimisation problems. It is shown that partnering strategies that exploit problem-specific knowledge are superior and can counter inappropriate (sub-) fitness measurements.

An Estimation of Distribution Algorithm for Nurse Scheduling

Schedules can be built in a similar way to a human scheduler by using a set of rules that involve domain knowledge. This paper presents an Estimation of Distribution Algorithm (EDA) for the nurse scheduling problem, which involves choosing a suitable scheduling rule from a set for the assignment of each nurse. Unlike previous work that used Genetic Algorithms (GAs) to implement implicit learning, the learning in the proposed algorithm is explicit, i.e. we identify and mix building blocks directly. The EDA is applied to implement such explicit learning by building a Bayesian network of the joint distribution of solutions. The conditional probability of each variable in the network is computed according to an initial set of promising solutions. Subsequently, each new instance for each variable is generated by using the corresponding conditional probabilities, until all variables have been generated, i.e. in our case, a new rule string has been obtained. Another set of rule strings will be generated in this way, some of which will replace previous strings based on fitness selection. If stopping conditions are not met, the conditional probabilities for all nodes in the Bayesian network are updated again using the current set of promising rule strings. Computational results from 52 real data instances demonstrate the success of this approach. It is also suggested that the learning mechanism in the proposed approach might be suitable for other scheduling problems.

Variation, correlation and heritability of interest characters for selection of African eggplant

The eggplant ( Solanum aethiopicum ) is the species of the Solanum genus, whose geographical distribution is broadest. It is grown throughout tropical Africa, and includes three groups of cultivars commonly called African or indigenous eggplant. Kumba group or “bitter eggplant” is an important Solanaceous vegetable crop in Burkina Faso. The objective of this study was to determine genetic variability, strength of association and level of heritability among agronomic interest traits. Phenotypic and genotypic variations and heritability of 14 traits were estimated in 61 accessions at Institut de Développement Rural (IDR), Gampela in Burkina Faso. High phenotypic and genotypic coefficients of variation were observed for fruit diameter, number of seeds per fruit, fruit weight, leaf blade length and width, and height at flowering. In addition, genetic and phenotypic variances were high for the number of seed, fruit weight, plant height at flowering and days to 50% flowering. High heritability estimates were recorded for all traits. Fruit weight showed a positive association with fruit diameter and thickness. The fifty percent flowering cycle registered positive correlations with plant height and fruit diameter. Fruit number showed a negative association with fruit weight and diameter, and 50% flowering cyle.

RESPONSE OF LOCALLY ADAPTED PEARL MILLET POPULATIONS TO S1 PROGENY RECURRENT SELECTION FOR GRAIN YIELD AND RESISTANCE TO RUST

In the semi-arid zones of Uganda, pearl millet ( Pennisetum glaucum (L.) R. Br.) is mainly grown for food and income; but rust (Puccinia substriata var indica (L.) R. Br.) is the main foliar constraint lowering yield. The objective of the study was to genetically improve grain yield and rust resistance of two locally adapted populations (Lam and Omoda), through two cycles of modified phenotypic S1 progeny recurrent selection. Treatments included three cycles of two locally adapted pearl millet populations, evaluated at three locations. Significant net genetic gain for grain yield (72 and 36%) were achieved in Lam and Omoda populations, respectively. This led to grain yield of 1,047 from 611 kg ha-1 in Lam population and 943 from 693 kg ha-1 in Omoda population. Significant improvement in rust resistance was achieved in the two populations, with a net genetic gain of -55 and -71% in Lam and Omoda populations, respectively. Rust severity reduced from 30 to 14% in Lam population and from 57 to 17% in Omoda population. Net positive genetic gains of 68 and 8% were also achieved for 1000-grain weight in Lam and Omoda, respectively. Traits with a net negative genetic gain in both populations were days to 50% flowering, days to 50% anthesis, days to 50% physiological maturity, flower-anthesis interval, plant height and leaf area.

Genetic algorithm seeding of idiotypic networks for mobile-robot navigation

Robot-control designers have begun to exploit the properties of the human immune system in order to produce dynamic systems that can adapt to complex, varying, real-world tasks. Jerne’s idiotypic-network theory has proved the most popular artificial-immune-system (AIS) method for incorporation into behaviour-based robotics, since idiotypic selection produces highly adaptive responses. However, previous efforts have mostly focused on evolving the network connections and have often worked with a single, preengineered set of behaviours, limiting variability. This paper describes a method for encoding behaviours as a variable set of attributes, and shows that when the encoding is used with a genetic algorithm (GA), multiple sets of diverse behaviours can develop naturally and rapidly, providing much greater scope for flexible behaviour-selection. The algorithm is tested extensively with a simulated e-puck robot that navigates around a maze by tracking colour. Results show that highly successful behaviour sets can be generated within about 25 minutes, and that much greater diversity can be obtained when multiple autonomous populations are used, rather than a single one.

A Bayesian optimization algorithm for the nurse scheduling problem

A Bayesian optimization algorithm for the nurse scheduling problem is presented, which involves choosing a suitable scheduling rule from a set for each nurse’s assignment. Unlike our previous work that used GAs to implement implicit learning, the learning in the proposed algorithm is explicit, i.e. eventually, we will be able to identify and mix building blocks directly. The Bayesian optimization algorithm is applied to implement such explicit learning by building a Bayesian network of the joint distribution of solutions. The conditional probability of each variable in the network is computed according to an initial set of promising solutions. Subsequently, each new instance for each variable is generated by using the corresponding conditional probabilities, until all variables have been generated, i.e. in our case, a new rule string has been obtained. Another set of rule strings will be generated in this way, some of which will replace previous strings based on fitness selection. If stopping conditions are not met, the conditional probabilities for all nodes in the Bayesian network are updated again using the current set of promising rule strings. Computational results from 52 real data instances demonstrate the success of this approach. It is also suggested that the learning mechanism in the proposed approach might be suitable for other scheduling problems.