'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,

A graph-based hyper heuristic for timetabling problems

This paper presents an investigation of a simple generic hyper-heuristic approach upon a set of widely used constructive heuristics (graph coloring heuristics) in timetabling. Within the hyperheuristic framework, a Tabu Search approach is employed to search for permutations of graph heuristics which are used for constructing timetables in exam and course timetabling problems. This underpins a multi-stage hyper-heuristic where the Tabu Search employs permutations upon a different number of graph heuristics in two stages. We study this graph-based hyper-heuristic approach within the context of exploring fundamental issues concerning the search space of the hyper-heuristic (the heuristic space) and the solution space. Such issues have not been addressed in other hyper-heuristic research. These approaches are tested on both exam and course benchmark timetabling problems and are compared with the fine-tuned bespoke state-of-the-art approaches. The results are within the range of the best results reported in the literature. The approach described here represents a significantly more generally applicable approach than the current state of the art in the literature. Future work will extend this hyper-heuristic framework by employing methodologies which are applicable on a wider range of timetabling and scheduling problems.

A graph-based hyper heuristic for timetabling problems

This paper presents an investigation of a simple generic hyper-heuristic approach upon a set of widely used constructive heuristics (graph coloring heuristics) in timetabling. Within the hyperheuristic framework, a Tabu Search approach is employed to search for permutations of graph heuristics which are used for constructing timetables in exam and course timetabling problems. This underpins a multi-stage hyper-heuristic where the Tabu Search employs permutations upon a different number of graph heuristics in two stages. We study this graph-based hyper-heuristic approach within the context of exploring fundamental issues concerning the search space of the hyper-heuristic (the heuristic space) and the solution space. Such issues have not been addressed in other hyper-heuristic research. These approaches are tested on both exam and course benchmark timetabling problems and are compared with the fine-tuned bespoke state-of-the-art approaches. The results are within the range of the best results reported in the literature. The approach described here represents a significantly more generally applicable approach than the current state of the art in the literature. Future work will extend this hyper-heuristic framework by employing methodologies which are applicable on a wider range of timetabling and scheduling problems.

Nurse Rostering with Genetic Algorithms

In recent years genetic algorithms have emerged as a useful tool for the heuristic solution of complex discrete optimisation problems. In particular there has been considerable interest in their use in tackling problems arising in the areas of scheduling and timetabling. However, the classical genetic algorithm paradigm is not well equipped to handle constraints and successful implementations usually require some sort of modification to enable the search to exploit problem specific knowledge in order to overcome this shortcoming. This paper is concerned with the development of a family of genetic algorithms for the solution of a nurse rostering problem at a major UK hospital. The hospital is made up of wards of up to 30 nurses. Each ward has its own group of nurses whose shifts have to be scheduled on a weekly basis. In addition to fulfilling the minimum demand for staff over three daily shifts, nurses’ wishes and qualifications have to be taken into account. The schedules must also be seen to be fair, in that unpopular shifts have to be spread evenly amongst all nurses, and other restrictions, such as team nursing and special conditions for senior staff, have to be satisfied. The basis of the family of genetic algorithms is a classical genetic algorithm consisting of n-point crossover, single-bit mutation and a rank-based selection. The solution space consists of all schedules in which each nurse works the required number of shifts, but the remaining constraints, both hard and soft, are relaxed and penalised in the fitness function. The talk will start with a detailed description of the problem and the initial implementation and will go on to highlight the shortcomings of such an approach, in terms of the key element of balancing feasibility, i.e. covering the demand and work regulations, and quality, as measured by the nurses’ preferences. A series of experiments involving parameter adaptation, niching, intelligent weights, delta coding, local hill climbing, migration and special selection rules will then be outlined and it will be shown how a series of these enhancements were able to eradicate these difficulties. Results based on several months’ real data will be used to measure the impact of each modification, and to show that the final algorithm is able to compete with a tabu search approach currently employed at the hospital. The talk will conclude with some observations as to the overall quality of this approach to this and similar problems.

'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,

An Indirect Genetic Algorithm for a Nurse Scheduling Problem

This paper describes a Genetic Algorithms approach to a manpower-scheduling problem arising at a major UK hospital. Although Genetic Algorithms have been successfully used for similar problems in the past, they always had to overcome the limitations of the classical Genetic Algorithms paradigm in handling the conflict between objectives and constraints. The approach taken here is to use an indirect coding based on permutations of the nurses, and a heuristic decoder that builds schedules from these permutations. Computational experiments based on 52 weeks of live data are used to evaluate three different decoders with varying levels of intelligence, and four well-known crossover operators. Results are further enhanced by introducing a hybrid crossover operator and by making use of simple bounds to reduce the size of the solution space. The results reveal that the proposed algorithm is able to find high quality solutions and is both faster and more flexible than a recently published Tabu Search approach.

A grouping model for distributed pipeline assets maintenance decision

Distributed pipeline assets systems are crucial to society. The deterioration of these assets and the optimal allocation of limited budget for their maintenance correspond to crucial challenges for water utility managers. Decision makers should be assisted with optimal solutions to select the best maintenance plan concerning available resources and management strategies. Much research effort has been dedicated to the development of optimal strategies for maintenance of water pipes. Most of the maintenance strategies are intended for scheduling individual water pipe. Consideration of optimal group scheduling replacement jobs for groups of pipes or other linear assets has so far not received much attention in literature. It is a common practice that replacement planners select two or three pipes manually with ambiguous criteria to group into one replacement job. This is obviously not the best solution for job grouping and may not be cost effective, especially when total cost can be up to multiple million dollars. In this paper, an optimal group scheduling scheme with three decision criteria for distributed pipeline assets maintenance decision is proposed. A Maintenance Grouping Optimization (MGO) model with multiple criteria is developed. An immediate challenge of such modeling is to deal with scalability of vast combinatorial solution space. To address this issue, a modified genetic algorithm is developed together with a Judgment Matrix. This Judgment Matrix is corresponding to various combinations of pipe replacement schedules. An industrial case study based on a section of a real water distribution network was conducted to test the new model. The results of the case study show that new schedule generated a significant cost reduction compared with the schedule without grouping pipes.

A natural model for architecture

In John Frazer's seminal book An Evolutionary Architecture (1995), from which this essay is extracted, a fundamental approach is established for have natural systems can unfold mechanisms for negotiating the complex design space inherent in architectural systems. In this essay, which forms a critical part of the book, Frazer draws both correlations and distinctions from natural processes as emulated in design processes and form as active manifestations within natural systems. Form is seen as an evolving agent generated via the rules of descriptive genetic coding, functioning as a part of a metabolic environment. Frazer's process-model establishes the realm in which computation must manoeuvre to produce a valid solution space, including the operations of self-organisation, complexity and emergent behaviour. Addressing design as an authored practice, he extends the transference of 'creativity' from the explicit impression into form, to the investment of though, organisation and strategy in the computational processes which produce form. Frazer's text concentrates astutely on the practising of the evolutionary paradigm, the output of which postulates an architecture born of the relationships to dynamic environmental and socio-economic contexts, and realised through morphogenetic materialisation.

Empower me : social innovation design for homeless families : collective design creativity

Design Creativity has largely been explored as an individual expression of design cognition rather than as the collective manifestation of interaction in context. Recent approaches to design with an emphasis on co-design suggest that the problem-solution space co-evolves through social interaction. Socially Responsive Design for Social Innovation constitutes the most recent and perhaps the most promising domain of application for design thinking practices that emphasize collaborative innovation. In this paper, we describe the ideation of a service design solution for homeless families (Em.power.me), developed through consultation with a range of stakeholders over a three month period. This service design innovation aimed to visualise how such a service would operate and identify the potential benefits for all stakeholders. We focus here on the phases leading to the ideation of the service design.

Efficiency of parallelisation of genetic algorithms in the data analysis context

Most real-life data analysis problems are difficult to solve using exact methods, due to the size of the datasets and the nature of the underlying mechanisms of the system under investigation. As datasets grow even larger, finding the balance between the quality of the approximation and the computing time of the heuristic becomes non-trivial. One solution is to consider parallel methods, and to use the increased computational power to perform a deeper exploration of the solution space in a similar time. It is, however, difficult to estimate a priori whether parallelisation will provide the expected improvement. In this paper we consider a well-known method, genetic algorithms, and evaluate on two distinct problem types the behaviour of the classic and parallel implementations.

An Algorithm to Decide Feasibility of Linear Integer Constraints Occurrng in Decision Tables

Abstract-To detect errors in decision tables one needs to decide whether a given set of constraints is feasible or not. This paper describes an algorithm to do so when the constraints are linear in variables that take only integer values. Decision tables with such constraints occur frequently in business data processing and in nonnumeric applications. The aim of the algorithm is to exploit. the abundance of very simple constraints that occur in typical decision table contexts. Essentially, the algorithm is a backtrack procedure where the the solution space is pruned by using the set of simple constrains. After some simplications, the simple constraints are captured in an acyclic directed graph with weighted edges. Further, only those partial vectors are considered from extension which can be extended to assignments that will at least satisfy the simple constraints. This is how pruning of the solution space is achieved. For every partial assignment considered, the graph representation of the simple constraints provides a lower bound for each variable which is not yet assigned a value. These lower bounds play a vital role in the algorithm and they are obtained in an efficient manner by updating older lower bounds. Our present algorithm also incorporates an idea by which it can be checked whether or not an (m - 2)-ary vector can be extended to a solution vector of m components, thereby backtracking is reduced by one component.

Equivalent constitutive model-based design of wave-absorbing material system and controller

A design methodology for wave-absorbing active material system is reported. The design enforces equivalence between an assumed material model having wave-absorbing behavior and a set of target feedback controllers for an array of microelectro-mechanical transducers which are integral part of the active material system. The proposed methodology is applicable to problems involving the control of acoustic waves in passive-active material system with complex constitutive behavior at different length-scales. A stress relaxation type one-dimensional constitutive model involving viscous damping mechanism is considered, which shows asymmetric wave dispersion characteristics about the half-line. The acoustic power flow and asymptotic stability of such material system are studied. A single sensor non-collocated linear feedback control system in a one-dimensional finite waveguide, which is a representative volume element in an active material system, is considered. Equivalence between the exact dynamic equilibrium of these two systems is imposed. It results in the solution space of the design variables, namely the equivalent damping coefficient, the wavelength(s) to be controlled and the location of the sensor. The characteristics of the controller transfer functions and their pole-placement problem are studied. (c) 2005 Elsevier Ltd. All rights reserved.

Binding of active site directed ligands to phospholipase A2: Implications on the molecular constraints and catalytic mechanism

Molecular constraints for the localization of active site directed ligands (competitive inhibitors and substrates) in the active site of phospholipase A2 (PLA2) are characterized. Structure activity relationships with known inhibitors suggest that the head : group interactions dominate the selectivity as well as a substantial part of the affinity. The ab initio fitting of the amide ligands in the active site was carried out to characterize the head group interactions. Based on a systematic coordinate space search, formamide is docked with known experimental constraints such as coordination of the carbonyl group to Ca2+ and hydrogen bond between amide nitrogen and ND1 of His48. An optimal position for a bound water molecule is identified and its significance for the catalytic mechanism is postulated. Unlike the traditional ''pseudo-triad'' mechanism, the ''Ca-coordinatedoxyanion'' mechanism proposed here invokes activation of the catalytic water to form the oxyanion in the coordination sphere of calcium. As it attacks the carbonyl carbon of the ester, a near-tetrahedral intermediate is formed. As the second proton of the catalytic water is abstracted by the ester oxygen, its reorientation and simultaneous cleavage form hydrogen bond with ND1 of His48. In this mechanism of esterolysis, a catalytic role for the water co-ordinated to Ca2+ is recognised.

An Efficient Robust Algorithm for the Surface-Potential Calculation of Independent DG MOSFET

Although the recently proposed single-implicit-equation-based input voltage equations (IVEs) for the independent double-gate (IDG) MOSFET promise faster computation time than the earlier proposed coupled-equations-based IVEs, it is not clear how those equations could be solved inside a circuit simulator as the conventional Newton-Raphson (NR)-based root finding method will not always converge due to the presence of discontinuity at the G-zero point (GZP) and nonremovable singularities in the trigonometric IVE. In this paper, we propose a unique algorithm to solve those IVEs, which combines the Ridders algorithm with the NR-based technique in order to provide assured convergence for any bias conditions. Studying the IDG MOSFET operation carefully, we apply an optimized initial guess to the NR component and a minimized solution space to the Ridders component in order to achieve rapid convergence, which is very important for circuit simulation. To reduce the computation budget further, we propose a new closed-form solution of the IVEs in the near vicinity of the GZP. The proposed algorithm is tested with different device parameters in the extended range of bias conditions and successfully implemented in a commercial circuit simulator through its Verilog-A interface.

Viscoelastic phenomenology based structure assignment for closed-loop vibration control of a beam with sensors and actuators

In this paper we incorporate a novel approach to synthesize a class of closed-loop feedback control, based on the variational structure assignment. Properties of a viscoelastic system are used to design an active feedback controller for an undamped structural system with distributed sensor, actuator and controller. Wave dispersion properties of onedimensional beam system have been studied. Efficiency of the chosen viscoelastic model in enhancing damping and stability properties of one-dimensional viscoelastic bar have been analyzed. The variational structure is projected on a solution space of a closed-loop system involving a weakly damped structure with distributed sensor and actuator with controller. These assign the phenomenology based internal strain rate damping parameter of a viscoelastic system to the usual elastic structure but with active control. In the formulation a model of cantilever beam with non-collocated actuator and sensor has been considered. The formulation leads to the matrix identification problem of two dynamic stiffness matrices. The method has been simplified to obtain control system gains for the free vibration control of a cantilever beam system with collocated actuator-sensor, using quadratic optimal control and pole-placement methods.