Physical model parameter optimisation for calibrated emulation of the Dallas Rangemaster Treble Booster guitar pedal

In this work we explore optimising parameters of a physical circuit model relative to input/output measurements, using the Dallas Rangemaster Treble Booster as a case study. A hybrid metaheuristic/gradient descent algorithm is implemented, where the initial parameter sets for the optimisation are informed by nominal values from schematics and datasheets. Sensitivity analysis is used to screen parameters, which informs a study of the optimisation algorithm against model complexity by fixing parameters. The results of the optimisation show a significant increase in the accuracy of model behaviour, but also highlight several key issues regarding the recovery of parameters.

Mesh partitioning: A multilevel balancing and refinement algorithm

Multilevel algorithms are a successful class of optimisation techniques which address the mesh partitioning problem. They usually combine a graph contraction algorithm together with a local optimisation method which refines the partition at each graph level. In this paper we present an enhancement of the technique which uses imbalance to achieve higher quality partitions. We also present a formulation of the Kernighan-Lin partition optimisation algorithm which incorporates load-balancing. The resulting algorithm is tested against a different but related state-of the-art partitioner and shown to provide improved results.

'Bayesian Optimisation for Nurse Scheduling'

A Bayesian optimisation algorithm for a nurse scheduling problem is presented, which involves choosing a suitable scheduling rule from a set for each nurse's assignment. When a human scheduler works, he normally builds a schedule systematically following a set of rules. After much practice, the scheduler gradually masters the knowledge of which solution parts go well with others. He can identify good parts and is aware of the solution quality even if the scheduling process is not yet completed, thus having the ability to finish a schedule by using flexible, rather than fixed, rules. In this paper, we design a more human-like scheduling algorithm, by using a Bayesian optimisation algorithm to implement explicit learning from past solutions. A nurse scheduling problem from a UK hospital is used for testing. Unlike our previous work that used Genetic Algorithms to implement implicit learning [1], the learning in the proposed algorithm is explicit, i.e. we identify and mix building blocks directly. The Bayesian optimisation 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, new rule strings have been obtained. Sets of rule strings are generated in this way, some of which will replace previous strings based on fitness. 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. For clarity, consider the following toy example of scheduling five nurses with two rules (1: random allocation, 2: allocate nurse to low-cost shifts). In the beginning of the search, the probabilities of choosing rule 1 or 2 for each nurse is equal, i.e. 50%. After a few iterations, due to the selection pressure and reinforcement learning, we experience two solution pathways: Because pure low-cost or random allocation produces low quality solutions, either rule 1 is used for the first 2-3 nurses and rule 2 on remainder or vice versa. In essence, Bayesian network learns 'use rule 2 after 2-3x using rule 1' or vice versa. It should be noted that for our and most other scheduling problems, the structure of the network model is known and all variables are fully observed. In this case, the goal of learning is to find the rule values that maximize the likelihood of the training data. Thus, learning can amount to 'counting' in the case of multinomial distributions. For our problem, we use our rules: Random, Cheapest Cost, Best Cover and Balance of Cost and Cover. In more detail, the steps of our Bayesian optimisation algorithm for nurse scheduling are: 1. Set t = 0, and generate an initial population P(0) at random; 2. Use roulette-wheel selection to choose a set of promising rule strings S(t) from P(t); 3. Compute conditional probabilities of each node according to this set of promising solutions; 4. Assign each nurse using roulette-wheel selection based on the rules' conditional probabilities. A set of new rule strings O(t) will be generated in this way; 5. Create a new population P(t+1) by replacing some rule strings from P(t) with O(t), and set t = t+1; 6. If the termination conditions are not met (we use 2000 generations), go to step 2. Computational results from 52 real data instances demonstrate the success of this approach. They also suggest that the learning mechanism in the proposed approach might be suitable for other scheduling problems. Another direction for further research is to see if there is a good constructing sequence for individual data instances, given a fixed nurse scheduling order. If so, the good patterns could be recognized and then extracted as new domain knowledge. Thus, by using this extracted knowledge, we can assign specific rules to the corresponding nurses beforehand, and only schedule the remaining nurses with all available rules, making it possible to reduce the solution space. Acknowledgements The work was funded by the UK Government's major funding agency, Engineering and Physical Sciences Research Council (EPSRC), under grand GR/R92899/01. References [1] Aickelin U, "An Indirect Genetic Algorithm for Set Covering Problems", Journal of the Operational Research Society, 53(10): 1118-1126,

'Bayesian Optimisation for Nurse Scheduling'

A Bayesian optimisation algorithm for a nurse scheduling problem is presented, which involves choosing a suitable scheduling rule from a set for each nurse's assignment. When a human scheduler works, he normally builds a schedule systematically following a set of rules. After much practice, the scheduler gradually masters the knowledge of which solution parts go well with others. He can identify good parts and is aware of the solution quality even if the scheduling process is not yet completed, thus having the ability to finish a schedule by using flexible, rather than fixed, rules. In this paper, we design a more human-like scheduling algorithm, by using a Bayesian optimisation algorithm to implement explicit learning from past solutions. A nurse scheduling problem from a UK hospital is used for testing. Unlike our previous work that used Genetic Algorithms to implement implicit learning [1], the learning in the proposed algorithm is explicit, i.e. we identify and mix building blocks directly. The Bayesian optimisation 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, new rule strings have been obtained. Sets of rule strings are generated in this way, some of which will replace previous strings based on fitness. 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. For clarity, consider the following toy example of scheduling five nurses with two rules (1: random allocation, 2: allocate nurse to low-cost shifts). In the beginning of the search, the probabilities of choosing rule 1 or 2 for each nurse is equal, i.e. 50%. After a few iterations, due to the selection pressure and reinforcement learning, we experience two solution pathways: Because pure low-cost or random allocation produces low quality solutions, either rule 1 is used for the first 2-3 nurses and rule 2 on remainder or vice versa. In essence, Bayesian network learns 'use rule 2 after 2-3x using rule 1' or vice versa. It should be noted that for our and most other scheduling problems, the structure of the network model is known and all variables are fully observed. In this case, the goal of learning is to find the rule values that maximize the likelihood of the training data. Thus, learning can amount to 'counting' in the case of multinomial distributions. For our problem, we use our rules: Random, Cheapest Cost, Best Cover and Balance of Cost and Cover. In more detail, the steps of our Bayesian optimisation algorithm for nurse scheduling are: 1. Set t = 0, and generate an initial population P(0) at random; 2. Use roulette-wheel selection to choose a set of promising rule strings S(t) from P(t); 3. Compute conditional probabilities of each node according to this set of promising solutions; 4. Assign each nurse using roulette-wheel selection based on the rules' conditional probabilities. A set of new rule strings O(t) will be generated in this way; 5. Create a new population P(t+1) by replacing some rule strings from P(t) with O(t), and set t = t+1; 6. If the termination conditions are not met (we use 2000 generations), go to step 2. Computational results from 52 real data instances demonstrate the success of this approach. They also suggest that the learning mechanism in the proposed approach might be suitable for other scheduling problems. Another direction for further research is to see if there is a good constructing sequence for individual data instances, given a fixed nurse scheduling order. If so, the good patterns could be recognized and then extracted as new domain knowledge. Thus, by using this extracted knowledge, we can assign specific rules to the corresponding nurses beforehand, and only schedule the remaining nurses with all available rules, making it possible to reduce the solution space. Acknowledgements The work was funded by the UK Government's major funding agency, Engineering and Physical Sciences Research Council (EPSRC), under grand GR/R92899/01. References [1] Aickelin U, "An Indirect Genetic Algorithm for Set Covering Problems", Journal of the Operational Research Society, 53(10): 1118-1126,

Automatic tuning method for the design of supplementary damping controllers for flexible alternating current transmission system devices

The design of supplementary damping controllers to mitigate the effects of electromechanical oscillations in power systems is a highly complex and time-consuming process, which requires a significant amount of knowledge from the part of the designer. In this study, the authors propose an automatic technique that takes the burden of tuning the controller parameters away from the power engineer and places it on the computer. Unlike other approaches that do the same based on robust control theories or evolutionary computing techniques, our proposed procedure uses an optimisation algorithm that works over a formulation of the classical tuning problem in terms of bilinear matrix inequalities. Using this formulation, it is possible to apply linear matrix inequality solvers to find a solution to the tuning problem via an iterative process, with the advantage that these solvers are widely available and have well-known convergence properties. The proposed algorithm is applied to tune the parameters of supplementary controllers for thyristor controlled series capacitors placed in the New England/New York benchmark test system, aiming at the improvement of the damping factor of inter-area modes, under several different operating conditions. The results of the linear analysis are validated by non-linear simulation and demonstrate the effectiveness of the proposed procedure.

Optimization techniques for power distribution planning with uncertainties: a comparative study

Important research effort has been devoted to the topic of optimal planning of distribution systems. The non linear nature of the system, the need to consider a large number of scenarios and the increasing necessity to deal with uncertainties make optimal planning in distribution systems a difficult task. Heuristic techniques approaches have been proposed to deal with these issues, overcoming some of the inherent difficulties of classic methodologies. This paper considers several methodologies used to address planning problems of electrical power distribution networks, namely mixedinteger linear programming (MILP), ant colony algorithms (AC), genetic algorithms (GA), tabu search (TS), branch exchange (BE), simulated annealing (SA) and the Bender´s decomposition deterministic non-linear optimization technique (BD). Adequacy of theses techniques to deal with uncertainties is discussed. The behaviour of each optimization technique is compared from the point of view of the obtained solution and of the methodology performance. The paper presents results of the application of these optimization techniques to a real case of a 10-kV electrical distribution system with 201 nodes that feeds an urban area.

Swarm intelligence for scheduling: a review

Swarm Intelligence generally refers to a problem-solving ability that emerges from the interaction of simple information-processing units. The concept of Swarm suggests multiplicity, distribution, stochasticity, randomness, and messiness. The concept of Intelligence suggests that problem-solving approach is successful considering learning, creativity, cognition capabilities. This paper introduces some of the theoretical foundations, the biological motivation and fundamental aspects of swarm intelligence based optimization techniques such Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO) and Artificial Bees Colony (ABC) algorithms for scheduling optimization.

A new MILP based approach for unit commitment in power production planning

This paper presents a complete, quadratic programming formulation of the standard thermal unit commitment problem in power generation planning, together with a novel iterative optimisation algorithm for its solution. The algorithm, based on a mixed-integer formulation of the problem, considers piecewise linear approximations of the quadratic fuel cost function that are dynamically updated in an iterative way, converging to the optimum; this avoids the requirement of resorting to quadratic programming, making the solution process much quicker. From extensive computational tests on a broad set of benchmark instances of this problem, the algorithm was found to be flexible and capable of easily incorporating different problem constraints. Indeed, it is able to tackle ramp constraints, which although very important in practice were rarely considered in previous publications. Most importantly, optimal solutions were obtained for several well-known benchmark instances, including instances of practical relevance, that are not yet known to have been solved to optimality. Computational experiments and their results showed that the method proposed is both simple and extremely effective.

A Hybrid Simulated Annealing approach to handle Energy Resource Management considering an intensive use of Electric Vehicles

The massification of electric vehicles (EVs) can have a significant impact on the power system, requiring a new approach for the energy resource management. The energy resource management has the objective to obtain the optimal scheduling of the available resources considering distributed generators, storage units, demand response and EVs. The large number of resources causes more complexity in the energy resource management, taking several hours to reach the optimal solution which requires a quick solution for the next day. Therefore, it is necessary to use adequate optimization techniques to determine the best solution in a reasonable amount of time. This paper presents a hybrid artificial intelligence technique to solve a complex energy resource management problem with a large number of resources, including EVs, connected to the electric network. The hybrid approach combines simulated annealing (SA) and ant colony optimization (ACO) techniques. The case study concerns different EVs penetration levels. Comparisons with a previous SA approach and a deterministic technique are also presented. For 2000 EVs scenario, the proposed hybrid approach found a solution better than the previous SA version, resulting in a cost reduction of 1.94%. For this scenario, the proposed approach is approximately 94 times faster than the deterministic approach.

Estado da Arte da Computação Evolutiva Aplicada à Eletrónica

A Computação Evolutiva enquadra-se na área da Inteligência Artificial e é um ramo das ciências da computação que tem vindo a ser aplicado na resolução de problemas em diversas áreas da Engenharia. Este trabalho apresenta o estado da arte da Computação Evolutiva, assim como algumas das suas aplicações no ramo da eletrónica, denominada Eletrónica Evolutiva (ou Hardware Evolutivo), enfatizando a síntese de circuitos digitais combinatórios. Em primeiro lugar apresenta-se a Inteligência Artificial, passando à Computação Evolutiva, nas suas principais vertentes: os Algoritmos Evolutivos baseados no processo da evolução das espécies de Charles Darwin e a Inteligência dos Enxames baseada no comportamento coletivo de alguns animais. No que diz respeito aos Algoritmos Evolutivos, descrevem-se as estratégias evolutivas, a programação genética, a programação evolutiva e com maior ênfase, os Algoritmos Genéticos. Em relação à Inteligência dos Enxames, descreve-se a otimização por colônia de formigas e a otimização por enxame de partículas. Em simultâneo realizou-se também um estudo da Eletrónica Evolutiva, explicando sucintamente algumas das áreas de aplicação, entre elas: a robótica, as FPGA, o roteamento de placas de circuito impresso, a síntese de circuitos digitais e analógicos, as telecomunicações e os controladores. A título de concretizar o estudo efetuado, apresenta-se um caso de estudo da aplicação dos algoritmos genéticos na síntese de circuitos digitais combinatórios, com base na análise e comparação de três referências de autores distintos. Com este estudo foi possível comparar, não só os resultados obtidos por cada um dos autores, mas também a forma como os algoritmos genéticos foram implementados, nomeadamente no que diz respeito aos parâmetros, operadores genéticos utilizados, função de avaliação, implementação em hardware e tipo de codificação do circuito.

Tunnel engineering – influence of the type and the quantity of measurements in the back analysis of geomechanical parameters

The monitoring data collected during tunnel excavation can be used in inverse analysis procedures in order to identify more realistic geomechanical parameters that can increase the knowledge about the interested formations. These more realistic parameters can be used in real time to adapt the project to the real structure in situ behaviour. However, monitoring plans are normally designed for safety assessment and not especially for the purpose of inverse analysis. In fact, there is a lack of knowledge about what types and quantity of measurements are needed to succeed in identifying the parameters of interest. Also, the optimisation algorithm chosen for the identification procedure may be important for this matter. In this work, this problem is addressed using a theoretical case with which a thorough parametric study was carried out using two optimisation algorithms based on different calculation paradigms, namely a conventional gradient-based algorithm and an evolution strategy algorithm. Calculations were carried for different sets of parameters to identify several combinations of types and amount of monitoring data. The results clearly show the high importance of the available monitoring data and the chosen algorithm for the success rate of the inverse analysis process.

Incorporating waiting time in competitive location models: formulations and heuristics

In this paper we propose a metaheuristic to solve a new version of the Maximum Capture Problem. In the original MCP, market capture is obtained by lower traveling distances or lower traveling time, in this new version not only the traveling time but also the waiting time will affect the market share. This problem is hard to solve using standard optimization techniques. Metaheuristics are shown to offer accurate results within acceptable computing times.

Incorporating waiting time in competitive location models: Formulations and heuristics

In this paper we propose a metaheuristic to solve a new version of the Maximum CaptureProblem. In the original MCP, market capture is obtained by lower traveling distances or lowertraveling time, in this new version not only the traveling time but also the waiting time willaffect the market share. This problem is hard to solve using standard optimization techniques.Metaheuristics are shown to offer accurate results within acceptable computing times.

Fitness and the level of homozygosity in a social insect.

To date very few studies have addressed the effects of inbreeding in social Hymenoptera, perhaps because the costs of inbreeding are generally considered marginal owing to male haploidy whereby recessive deleterious alleles are strongly exposed to selection in males. Here, we present one of the first studies on the effects of queen and worker homozygosity on colony performance. In a wild population of the ant Formica exsecta, the relative investment of single-queen colonies in sexual production decreased with increased worker homozygosity. This may either stem from increased homozygosity decreasing the likelihood of diploid brood to develop into queens or a lower efficiency of more homozygous workers at feeding larvae and thus a lower proportion of the female brood developing into queens. There was also a significant negative association between colony age and the level of queen but not worker homozygosity. This association may stem from inbreeding affecting queen lifespan and/or their fecundity, and thus colony survival. However, there was no association between queen homozygosity and colony size, suggesting that inbreeding affects colony survival as a result of inbred queens having a shorter lifespan rather than a lower fecundity. Finally, there was no significant association between either worker or queen homozygosity and the probability of successful colony founding, colony size and colony productivity, the three other traits studied. Overall, these results indicate that inbreeding depression may have important effects on colony fitness by affecting both the parental (queen) and offspring (worker)generations cohabiting within an ant colony.

Current status of a model system: the gene Gp-9 and its association with social organization in fire ants.

The Gp-9 gene in fire ants represents an important model system for studying the evolution of social organization in insects as well as a rich source of information relevant to other major evolutionary topics. An important feature of this system is that polymorphism in social organization is completely associated with allelic variation at Gp-9, such that single-queen colonies (monogyne form) include only inhabitants bearing B-like alleles while multiple-queen colonies (polygyne form) additionally include inhabitants bearing b-like alleles. A recent study of this system by Leal and Ishida (2008) made two major claims, the validity and significance of which we examine here. After reviewing existing literature, analyzing the methods and results of Leal and Ishida (2008), and generating new data from one of their study sites, we conclude that their claim that polygyny can occur in Solenopsis invicta in the U.S.A. in the absence of expression of the b-like allele Gp-9(b) is unfounded. Moreover, we argue that available information on insect OBPs (the family of proteins to which GP-9 belongs), on the evolutionary/population genetics of Gp-9, and on pheromonal/behavioral control of fire ant colony queen number fails to support their view that GP-9 plays no role in the chemosensory-mediated communication that underpins regulation of social organization. Our analyses lead us to conclude that there are no new reasons to question the existing consensus view of the Gp-9 system outlined in Gotzek and Ross (2007).