Search

Meta-heuristics are commonly used to solve combinatorial problems in practice. Many approaches provide very good quality solutions in a short amount of computational time; however most meta-heuristics use parameters to tune the performance of the meta-heuristic for particular problems and the selection of these parameters before solving the problem can require much time. This paper investigates the problem of setting parameters using a typical meta-heuristic called Meta-RaPS (Metaheuristic for Randomized Priority Search.). Meta-RaPS is a promising meta-heuristic optimization method that has been applied to different types of combinatorial optimization problems and achieved very good performance compared to other meta-heuristic techniques. To solve a combinatorial problem, Meta-RaPS uses two well-defined stages at each iteration: construction and local search. After a number of iterations, the best solution is reported. Meta-RaPS performance depends on the fine tuning of two main parameters, priority percentage and restriction percentage, which are used during the construction stage. This paper presents two different dynamic parameter setting methods for Meta-RaPS. These dynamic parameter setting approaches tune the parameters while a solution is being found. To compare these two approaches, nonparametric statistic approaches are utilized since the solutions are not normally distributed. Results from both these dynamic parameter setting methods are reported.

This work studies a scheduling problem where each job must be either accepted and scheduled to complete within its specified due window, or rejected altogether. Each job has a certain processing time and contributes a certain profit if accepted or penalty cost if rejected. There is a set of renewable resources, and no resource limit can be exceeded at any time. Each job requires a certain amount of each resource when processed, and the objective is to maximize total profit. A mixed-integer programming formulation and three approximation algorithms are presented: a priority rule heuristic, an algorithm based on the metaheuristic for randomized priority search and an evolutionary algorithm. Computational experiments comparing these four solution methods were performed on a set of generated benchmark problems covering a wide range of problem characteristics. The evolutionary algorithm outperformed the other methods in most cases, often significantly, and never significantly underperformed any method. [ABSTRACT FROM PUBLISHER]

We examine the winner determination problem for combinatorial auctions with sub-cardinality constraints (WDP-SC). In this type of auction, bidders submit bids for packages of items of interest together with a specific number of items they want. All items in a package are equally acceptable, but each bidder wants only the specified number of items in their package and not necessarily all. This type of auction is particularly relevant in cases where bidders are interested in items in close proximity to one another and view items as largely substitutable. We show that WDP-SC is NP-hard. We then develop an integer programming (IP) formulation and two approximation algorithms for solving WDP-SC: one based on simulated annealing (SA) and one based on the Metaheuristic for Randomized Priority Search (Meta-RaPS). We assess the performance of these three methods through computational tests performed on sets of large benchmark problems (ranging from 1000 to 10,000 bids) generated from four different distributions: two developed specifically to simulate proximity concerns, and two others previously used in the literature. IP produced the best solutions for 1000-bid problems, but in many cases SA and Meta-RaPS produced better solutions than IP on larger problems. Both IP and Meta-RaPS appear to be robust solution methods for this problem and consistently produced high-quality solutions across all problem sizes and distributions. [ABSTRACT FROM AUTHOR]

A single machine that includes loading/unloading areas for each job processed, loading and unloading could be performed while the machine is running causing minimized jobs completion time with lowest machine idle time. This design requires special kind of scheduling technique to ensure the accomplishment of those objectives if jobs’ processing, loading and unloading times are varying. The machine is modelled as a flow shop with blocking constraint. This research focuses on finding a solution to schedule this special case of flow shop of more than two machines with objective of minimizing jobs maximum completion time (makespan). The proposed solution in this research includes using a newly developed meta-heuristic known as Meta-heuristic for Randomized Priority Search (Meta-RaPS). Meta-RaPS construction phase is applied with the use of NEH flow shop scheduling algorithm and would provide very good schedules. The suggested technique is evaluated in comparison to top performing current meta-heuristics and construction heuristics on the famous benchmark flow shop data set of Taillard 20 . The results would suggest that applying Meta-RaPS for this flow shop problem is a great choice for constructing solutions and would provide high quality solutions with opportunity of more improvement in further research. [ABSTRACT FROM AUTHOR]

Though metaheuristics have been frequently employed to improve the performance of data mining algorithms, the opposite is not true. This paper discusses the process of employing a data mining algorithm to improve the performance of a metaheuristic algorithm. The targeted algorithms to be hybridized are the Meta-heuristic for Randomized Priority Search (Meta-RaPS) and an algorithm used to create an Inductive Decision Tree. This hybridization focuses on using a decision tree to perform on-line tuning of the parameters in Meta-RaPS. The process makes use of the information collected during the iterative construction and improvement phases Meta-RaPS performs. The data mining algorithm is used to find a favorable parameter using the knowledge gained from previous Meta-RaPS iterations. This knowledge is then used in future Meta-RaPS iterations. The proposed concept is applied to benchmark instances of the Vehicle Routing Problem. [ABSTRACT FROM AUTHOR]

Abstract: In their search for satisfactory solutions to complex combinatorial problems, metaheuristics methods are expected to intelligently explore the solution space. Various forms of memory have been used to achieve this goal and improve the performance of metaheuristics, which warranted the development of the Adaptive Memory Programming (AMP) framework [1]. This paper follows this framework by integrating Machine Learning (ML) concepts into metaheuristics as a way to guide metaheuristics while searching for solutions. The target metaheuristic method is Meta-heuristic for Randomized Priority Search (Meta-RaPS). Similar to most metaheuristics, Meta-RaPS consists of construction and improvement phases. Randomness coupled with a greedy heuristic are typically employed in the construction phase. While a local search heuristic is used in the improvement phase. This paper proposes adding a new learning phase, in which a ML method will be integrated. An Inductive Decision Tree (IDT) will be incorporated into the learning phase in an effort to learn from the information collected during the construction and improvement phases. The proposed approach will be demonstrated using instances for the Capacitated Vehicle Routing Problem (CVRP). [Copyright &y& Elsevier]

This article addresses the aerial refuelling scheduling problem (ARSP), where a set of fighter jets (jobs) with certain ready times must be refuelled from tankers (machines) by their due dates; otherwise, they reach a low fuel level (deadline) incurring a high cost. ARSP is an identical parallel machine scheduling problem with release times and due date-to-deadline windows to minimize the total weighted tardiness. A simulated annealing (SA) and metaheuristic for randomized priority search (Meta-RaPS) with the newly introduced composite dispatching rule, apparent piecewise tardiness cost with ready times (APTCR), are applied to the problem. Computational experiments compared the algorithms’ solutions to optimal solutions for small problems and to each other for larger problems. To obtain optimal solutions, a mixed integer program with a piecewise weighted tardiness objective function was solved for up to 12 jobs. The results show that Meta-RaPS performs better in terms of average relative error but SA is more efficient. [ABSTRACT FROM PUBLISHER]