Abdullah Alsheddy

Guided Local Search

Combinatorial explosion is a well-known phenomenon that prevents complete algorithms from solving many real-life combinatorial optimization problems. In many situations, heuristic search methods are needed. This chapter describes the principles of Guided Local Search (GLS) and Fast Local Search (FLS) and surveys their applications. GLS is a penalty-based metaheuristic algorithm that sits on top of other local search algorithms, with the aim to improve their efficiency and robustness. FLS is a way of reducing the size of the neighbourhood to improve the efficiency of local search. The chapter also provides guidance for implementing and using GLS and FLS. Four problems, representative of general application categories, are examined with detailed information provided on how to build a GLS-based method in each case.

Empowerment scheduling for a field workforce

Employee empowerment is a flexible management concept. As in traditional scheduling, the employer is still in charge of assigning jobs to staff. However, employees are allowed to express their preferences for the jobs they want to do. The hope is that empowerment will improve morale, which will improve productivity. The challenge is to design such an empowerment scheduling system without undesirable outcomes.In the proposed model, employees submit their preferences as “work plans”. The organizational goal and the employees’ work plans may not be in conflict. In such situations, win-win schedules can be generated without costing the organization. When there is a conflict, the organization is willing to give up a certain amount of its optimality (which is determined by the organization) in order to consider the employee’s work plans. The employer is in charge, and therefore jobs undesirable to any of the employees will still be done. A main consideration in empowerment is to make the employees feel that the system is fair. The proposed model maintains fairness by incorporating an automatic market-like mechanism that controls the violation cost of each employee’s request.The model is applied to solve a workforce scheduling problem which involves scheduling a multi-skilled workforce to geographically dispersed tasks. Extensive computational experiments are conducted, which show that this model enables an organization to implement employee empowerment effectively.

Guided Pareto Local Search based frameworks for biobjective optimization

Guided Pareto Local Search (GPLS) is an extension to the Guided Local Search algorithm to contain multiobjective combinatorial optimization. GPLS is shown to improve the convergence of the underlying Pareto local search algorithms. This paper demonstrates the potential of GPLS to be an effective searching technique that can be a central part of a multi-phase or hybrid frameworks. To confirm this, two simple frameworks based on GPLS are proposed: iGPLS and mGPLS. Both frameworks only require an initial set of diverse solutions. While GPLS starts from a randomly (or heuristically) generated solution, iGPLS starts with the initial diverse solution set. On the other hand, mGPLS is a parallel version of GPLS, in which each GPLS run starts independently from a solution in the initial set. The application of these frameworks to the biobjective 0/1 knapsack problem reveals the effectiveness of the GPLS based frameworks, demonstrated by achieving state-of-the-art results.

On the partitioning of dynamic workforce scheduling problems

This problem is based on the British Telecom workforce scheduling problem, in which technicians (with different skills) are assigned to tasks (which require different skills) which arrive (partially) dynamically during the day. In order to manage their workforce, British Telecom divides the different regions into several areas. At the beginning of each day all the technicians in a region are assigned to one of these areas. During the day, each technician is limited to tasks within the assigned area.This effectively decomposes a large dynamic scheduling problem into smaller problems. On one hand, it makes the problem more manageable. On the other hand, it gives rise to, potentially, a mismatch between technicians and tasks within an area. Furthermore, it prevents technicians from being assigned a job which is just outside their area but happens to be close to where they are currently working.This paper studies the effect of the number of partitions on the expected objective (number of completed tasks) that a rule-based system (responsible for the dynamic assignment and reassignment of tasks to resources following dynamic events) can reach.

On the investigation of hyper-heuristics on a financial forecasting problem

Financial forecasting is a really important area in computational finance, with numerous works in the literature. This importance can be reflected in the literature by the continuous development of new algorithms. Hyper-heuristics have been successfully used in the past for a number of search and optimization problems, and have shown very promising results. To the best of our knowledge, they have not been used for financial forecasting. In this paper we present pioneer work, where we use different hyper-heuristics frameworks to investigate whether we can improve the performance of a financial forecasting tool called EDDIE 8. EDDIE 8 allows the GP (Genetic Programming) to search in the search space of indicators for solutions, instead of using pre-specified ones; as a result, its search area has dramatically increased and sometimes solutions can be missed due to ineffective search. We apply 14 different low-level heuristics to EDDIE 8, to 30 different datasets, and examine their effect to the algorithm’s performance. We then select the most prominent heuristics and combine them into three different hyper-heuristics frameworks. Results show that all three frameworks are competitive, and are able to show significantly improved results, especially in the case of best results. Lastly, analysis on the weights of the heuristics shows that there can be a constant swinging among some of the low-level heuristics, which denotes that the hyper-heuristics frameworks are able to ‘know’ the appropriate time to switch from one heuristic to the other, based on their effectiveness.

On the partitioning of dynamic scheduling problems -: assigning technicians to areas

BT workforce scheduling problem considers technicians (with different skills) which are assigned to tasks which arrive (partially) dynamically during the day. In order to manage their workforce, BT divides the different regions into several areas. In the beginning of each day all the technicians in a region are assigned to one of these areas. During the day, tasks can only be allocated to technicians from the same area. In this paper we use a (1+1) EA in order to decide, once the area have been defined, which technicians to assign to which areas.

Off-line parameter tuning for Guided Local Search using Genetic Programming

Guided Local Search (GLS), which is a simple meta-heuristic with many successful applications, has lambda as the only parameter to tune. There has been no attempt to automatically tune this parameter, resulting in a parameterless GLS. Such a result is a very practical objective to facilitate the use of meta-heuristics for end- users (e.g. practitioners and researchers). In this paper, we propose a novel parameter tuning approach by using Genetic Programming (GP). GP is employed to evolve an optimal formula that GLS can use to dynamically compute lambda as a function of instance-dependent characteristics. Computational experiments on the travelling salesman problem demonstrate the feasibility and effectiveness of this approach, producing parameterless formulae with which the performance of GLS is competitive (if not better) than the standard GLS.

A penalty-based multi-objectivization approach for single objective optimization

Abstract Advances in Pareto optimization techniques have encouraged the study of their application to solve single-objective optimization problems. The motive is that the Pareto concept can be an effective approach to reduce the impacts of local optima. The most challenging task in developing such an approach is the reformulation of the target single-objective to multiple objectives. This paper proposes a new multi-objectivization approach by introducing an additional helper objective that is to be optimized with the primary objective simultaneously using Pareto local search. As a key feature, the additional objective is formulated as a function of the primary objective and penalties associated to solution features. The penalties are dynamically updated during the search, with the hope to guide the search to avoid non-promising features for the primary objective. Computational results on the traveling salesman problem and the quadratic assignment problem confirm the effectiveness of the proposed approach in comparison to other multi-objectivization approaches and state-of-the-art methods on these benchmarks.

The degree of dynamism for workforce scheduling problem with stochastic task duration

Real time dispatching strategies in a dynamic environment is a growing area of interest. Most of current work focuses mainly on two dynamic aspects of the problem, namely dynamic arrival of jobs and dynamic travel time. The degree of dynamism, for example is defined with respect to dynamic arrival of jobs. This paper focuses on another dynamic aspect, namely the duration of tasks. This aspect becomes important when tasks durations are relatively long and, in addition, one has to respect time windows. We characterize the degree of dynamism of such problems and show that it relates with the expected cost of a static scheduler which is reapplied in light of dynamic events. Furthermore, preliminary experiments indicate that the performance of the scheduler can be improved when the expected duration of a task is overestimated.