Showing papers on "Heuristic (computer science) published in 1993"

Dynamic variable ordering for ordered binary decision diagrams

Abstract: The ordered binary decision diagram (OBDD) has proven useful in many applications as an efficient data structure for representing and manipulating Boolean functions. A serious drawback of OBDD's is the need for application-specific heuristic algorithms to order the variables before processing. Further, for many problem instances in logic synthesis, the heuristic ordering algorithms which have been proposed are insufficient to allow OBDD operations to complete within a limited amount of memory. The paper proposes a solution to these problems based on having the OBDD package itself determine and maintain the variable order. This is done by periodically applying a minimization algorithm to reorder the variables of the OBDD to reduce its size. A new OBDD minimization algorithm, called the sifting algorithm, is proposed and appears especially effective in reducing the size of the OBDD. Experiments with dynamic variable ordering on the problem of forming the OBDD's for the primary outputs of a combinational circuit show that many computations complete using dynamic variable ordering when the same computation fails otherwise.

Design-to-time real-time scheduling

Abstract: Design-to-time is an approach to problem solving in resource-constrained domains where multiple solution methods are available for tasks. Those solution methods make tradeoffs in solution quality within the available time. The design-to-time approach is defined in detail, and contrasted to the anytime algorithm approach. A heuristic algorithm for design-to-time real-time scheduling is presented. The authors' blackboard architecture that implements the design-to-time approach is discussed, and an example problem and solution from the Distributed Vehicle Monitoring Testbed (DVMT) is described in detail. Experimental results, generated using simulation, show the effects of various parameters on scheduler performance. Future research goals and plans are discussed. >

Solving Optimally the Static Ground-Holding Policy Problem in Air Traffic Control

Abstract: As air traffic congestion grows, ground-holding (or “gate-holding”) of aircraft is becoming increasingly common. The “ground-holding policy problem” (GHPP) consists of developing strategies for deciding which aircraft to hold on the ground and for how long. In this paper we present a stochastic linear programming solution to the static GHPP for a single airport. The computational complexity of existing solutions requires heuristic approaches in order to solve practical instances of the problem. The advantage of our solution is that, even for the largest airports, problem instances result in linear programs that can be solved optimally using just a personal computer. We present a set of algorithms and compare their performance to a deterministic solution and to the passive strategy of no ground-holds (i.e., to the strategy of taking all delays in the air) under different weather scenarios.

Demand driven dispatch: a method for dynamic aircraft capacity assignment, models, and algorithms

Abstract: A major problem for the airline industry is the assignment of airplane capacity to flight schedules to meet fluctuating market needs. Demand Driven Dispatch (D3) is an operating concept that addresses this problem. Utilizing a demand forecast which improves as flight departure approaches, aircraft are dynamically assigned to flights to better match the predicted final demands. The result, demonstrated in studies of actual airline systems, is an increase in passenger loads and revenues with simultaneously reduced costs for a net of 1–5% improvement in operating profits. Concept implementation is simplified by the prevalence of yield management systems which provide the forecasting capability, and the emergence of airplane families which provide the necessary operational flexibility. Implementation also requires frequent solution of extremely large aircraft assignment problems. These problems, which can be formulated in terms of a multicommodity network flow, can be solved with heuristic algorithms shown to...

Strategic flow management for air traffic control

Abstract: One of the most important functions of air traffic management systems is the assignment of ground-holding times to flights, i.e., the determination of whether and by how much the take-off of a particular aircraft headed for a congested part of the ATC system should be postponed to reduce the likelihood and extent of airborne delays. In this paper, we will present an analysis of the fundamental case in which flights from many origins must be scheduled for arrival at a single, congested airport. We will describe a set of approaches for addressing a deterministic and a stochastic version of the problem. A minimum cost flow algorithm can be used for the deterministic problem. Under a particular natural assumption regarding the functional form of delay costs, a very efficient, simple algorithm is also available. For the stochastic version, an exact dynamic programming formulation turns out to be impractical for typical instances of the problem and we present a number of heuristic approaches to it. The models a...

Perfect hash functions made parallel-Lazy functional programming on a distributed multiprocessor

Abstract: A programming technique for efficient parallel search is described. The authors study a search problem for which a heuristic preprocess makes sequential execution feasible. Two key questions are addressed. (1) How can this algorithm, optimized for sequential execution, be programmed in parallel to produce significant speedup? (2) how can this be done in a purely functional language without compromising either conciseness or referential transparency? The authors describe programming techniques for efficient parallel search in a lazy and pure functional language. These techniques are applied to an illustrative example. Results of execution on a real parallel machine are given. >

Solving the maximum clique problem using a tabu search approach

Abstract: We describe two variants of a tabu search heuristic, a deterministic one and a probabilistic one, for the maximum clique problem. This heuristic may be viewed as a natural alternative implementation of tabu search for this problem when compared to existing ones. We also present a new random graph generator, the\(\hat p\)-generator, which produces graphs with larger clique sizes than comparable ones obtained by classical random graph generating techniques. Computational results on a large set of test problems randomly generated with this new generator are reported and compared with those of other approximate methods.

Heuristic algorithms for the multiple depot vehicle scheduling problem

Abstract: We consider the NP-hard Multiple Depot Vehicle Scheduling Problem, in which a given set of time-tabled trips have to be assigned to vehicles stationed at different depots, so as to minimize the number of vehicles used and the overall operational cost. The problem arises in the management of transportation companies. In this paper some structural properties of the problem are studied and used to design a new polynomial-time heuristic algorithm which always guarantees the use of the minimum number of vehicles. Several effective refining procedures are also proposed. Extensive computational results on test problems involving up to 1,000 trips and 10 depots are reported, showing that the new approach always produces very tight approximate solutions in small computing times and outperforms other heuristics from the literature.

Massively parallel tabu search for the quadratic assignment problem

Abstract: A new heuristic algorithm to perform tabu search on the Quadratic Assignment Problem (QAP) is developed. A massively parallel implementation of the algorithm on the Connection Machine CM-2 is provided. The implementation usesn2 processors, wheren is the size of the problem. The elements of the algorithm, calledPar_tabu, include dynamically changing tabu list sizes, aspiration criterion and long term memory. A new intensification strategy based on intermediate term memory is proposed and shown to be promising especially while solving large QAPs. The combination of all these elements gives a very efficient heuristic for the QAP: the best known or improved solutions are obtained in a significantly smaller number of iterations than in other comparative studies. Combined with the implementation on CM-2, this approach provides suboptimal solutions to QAPs of bigger dimensions in reasonable time.

A heuristic for the dynamic facility layout problem

Abstract: This paper presents a heuristic for the dynamic facility layout problem. All existing methods for solving this problem require the use of a dynamic programming model, the optimal solution of the quadratic assignment problem, or both. The proposed heuristic is based on the steepest-descent pairwise-interchange procedure to develop layouts utilizing material handling cost data from varying lengths of forecast windows as well as the explicit consideration of the corresponding rearrangement costs. Numeric results indicate that it typically performs as well as any existing methodology and only slightly worse than optimal.

Performance-Driven Steiner Tree Algorithms for Global Routing

Abstract: This paper presents two performance-driven Steiner tree algorithms for global routing which consider the minimization of timing delay during the tree construction as the goal One algorithm is based on nonlinear optimization method, another uses heuristic approach to guide the construction of Steiner tree A new timing model is established which includes both total length and critical path between source and sink in delay formulation, and an upper bound for timing delay is deducted and used to guide both algorithms Experiment results are given to demonstrate the effectiveness of the two algorithms

A combined optimization and simulation approach for designing automated storage/retrieval systems

Abstract: The study of Automated Storage/Retrieval Systems (AS/RS) in warehouses has developed along two main lines: One seeks to minimize total cost of an AS/RS, while the other explores the dynamic behavior of such a system. This study addressing the two issues simultaneously, obtains design parameters for a system that complies with desired levels of performance. To this end, a heuristic recursive optimization/simulation procedure is developed. It is also assumed that the number of stacker cranes (an important cost component of AS/RS) can be less than or equal to the number of aisles. The proposed procedure was applied to several situations, and converged within a few iterations.

A facility layout method for flexible manufacturing systems

Abstract: This facility layout of a flexible manufacturing system (FMS) involves the positioning of cells within given boundaries, so as to minimize the total projected travel time between cells. Defining the layout includes specifying the spatial coordinates of each cell, its orientation in either a horizontal or vertical position, and the location of its load/unload point. We refer to this problem as the FMS facility layout problem (FLP). In this paper we present a four-step heuristic methodology for solving the FLP. This heuristic combines variable partitioning and integer programming methods to generate an open field type of layout.

Integrating target analysis and tabu search for improved scheduling systems

Abstract: This paper explores the integration of the Artificial Intelligence/Operations Research approach known as target analysis with tabu search to create a more effective system for machine scheduling. Target analysis is designed to give heuristic and optimization procedures the ability to learn what rules are best for solving particular classes of problems. The authors focus on the development of rules that depend on memory functions to incorporate diversifying elements in a tabu search method which is tailored to find optimal or near optimal solutions for a class of single machine scheduling problems with delay penalties and setup costs.

Asynchronous organizations for multi-algorithm problems

Abstract: A Multi-Algorithm Problem (MAP) is a problem with many approximate algorithms available to solve it. Examples of MAPs are most of the combinatorial optimization problems and the multi-criterion problems with conflicting objectives. Since a MAP has no single algorithm able to solve it optimally in reasonable time, and each available algorithm for a MAP has its own approach to solving a problem, there exists the possibility of combining these algorithms in order to take advantage of their strengths. This thesis demonstrates that by using an Asynchronous Team, a software organization characterized by cyclic, iterative data flow and autonomous agents which communicate asynchronously through shared memories, one can synergistically combine many algorithms in order to reach better results than each algorithm can do by itself. As an example of a MAP, we chose the Euclidean version of the Traveling Salesman Problem (TSP), which is a difficult combinatorial problem and has many heuristic algorithms that only provide approximate solutions for it. Asynchronous Teams with a few algorithms were able to reach optimal solutions for all the TSP instances tackled whereas, individual algorithms could not. Moreover, parallel execution of Asynchronous Teams on a computer network presented linear speed up. We also observed that Asynchronous Teams for solving TSP are scale efficient; that is, the more algorithms an Asynchronous Team uses, the better it performs. We tested and analyzed several design parameters for Asynchronous Teams such as selection and destruction policies (how agents select and destroy data from shared memories,) initialization policies (how to initiate shared memories,) memory sizes, and the influence of data flows on the performance of Asynchronous Teams. We also present some specially developed algorithms to be exclusively executed in Asynchronous Teams. Finally, we present a Markov-based model that not only explains the behavior, but also helps in designing Asynchronous Teams by giving insights about expected value of final solutions, optimal relative execution frequencies of algorithms, and convergence of Asynchronous Teams.

Cost Minimization of Partitions into Multiple Devices

Abstract: This paper considers the problem of obtaining a minimum-cost partitioning of a large logic circuit into a collection of subcircuits implementable with devices selected from a given library. Each device in the library may have a different price, size, and terminal capacity. We propose a multi-way partitioning algorithm based on a recursive application of the Fiduccia-Mattheyses bipartitioning heuristic, extended to handle (a) the overall goal of the cost minimization and (b) the constraints reflecting the limitations on the capacity of FPGA chips. The experimental implementation of the proposed algorithm has exhibited a very encouraging performance, producing solutions close to the theoretical minima calculated for many benchmark circuits.

Hybrid genetic algorithm to optimize signal phasing and timing

Abstract: Signal timing optimization involves the selection of four basic design elements: phase sequence, cycle length, green split, and offset. None of the available signal timing models is considered adequate to optimize all four design elements, particularly in two-dimensional networks. Among the current models, TRANSYT-7F is most effective for timing, but it does not optimize phasing. Researchers have considered several methods for enhancing TRANSYT-7F to include phasing optimization but thus far no method has proven practical. An exhaustive search of possible phasing combinations is computationally prohibitive; thus a new approach is needed. Genetic algorithms (GAs) are heuristic probabilistic search procedures that have been applied to a wide range of engineering problems. The use is investigated of a GA in combination with the TRANSYT-7F optimization routine to select all signal timing design elements. The main purpose of the GA in the proposed scheme is to optimize phase sequences. Two implementations of the GA model are presented. In the first, the GA and TRANSYT-7F optimization routines are executed concurrently to achieve an optimal solution. In the second, the GA is allowed to optimize cycle length, phase sequences, and offsets. Then TRANSTY-7F is used to adjust the resultant signal timing. The results suggest that both implementations have potential for optimizing signal phasing and timing. However, the first method produces more consistent results. It also requires longer execution time.

An introduction to population approaches for optimization and hierarchical objective functions: a discussion on the role of tabu search

Abstract: Population approaches suitable for global combinatorial optimization are discussed in this paper. They are composed of a number of distinguishable individuals called "agents", each one using a particular optimization strategy. Periods of independent search follow phases on which the population is restarted from new configurations. Due to its intrinsic parallelism and the asynchronicity of the method, it is particularly suitable for parallel computers. Results on two test problems are presented in this paper. The individual search optimization strategies for each agent have been chosen having the basic characteristics of tabu search. This has been done in order to avoid mixing the hypothesized properties of these population approaches with those of more elaborate tabu search strategies, but remarking on its main characteristics. A set of four test problem "landscapes" is discussed and their use to improve and benchmark the results by using tabu search as the individual optimization strategy within a population heuristic is suggested and explored. The application of tabu search to new problem areas, like molecular biology, is also investigated.

Heuristic Procedures for Practical‐Sized Uncapacitated Location‐Capacitated Routing Problems*

Abstract: Facility location and vehicle routing are two important logistical problems closely interrelated in many real-world applications where locating facilities and determining the associated multi-stop vehicle routes are required simultaneously. Previous research has found that using the classical facility location models on these location-routing problems (LRPs) may lead to suboptimal solutions. We propose an approximate approach for the LRPs, which first generates and improves feasible location/allocation schemes with the associated multi-stop routing costs approximated using some route length estimators. We then design the minimum-cost routes based on the location/allocation results. We review two estimators that can provide accurate approximations to the multi-stop route distances; define the uncapacitated location-capacitated routing problem; and evaluate several heuristic procedures for approximately solving the problem. Computational results show that when vehicle capacities are not too restrictive, the sequential procedures that incorporate the two robust route length estimators can produce good solutions to practical-sized problems with a reasonable amount of computational efforts.

A continuous based heuristic for the maximum clique problem.

Abstract: We develop a deterministic heuristic based on a continuous quadratic nonconvex formulation of the maximum clique problem. The developed heuristic avoids the need for the calibration of parameters while maintaining computational competitiveness to greedy randomized search procedures which require such calibration. Furthermore it demonstrates the potential for using continuous approaches to model discrete problems. Experimental results on test problems.

A coordinated control of voltage and reactive power by heuristic modeling and approximate reasoning

Abstract: A flexible approximate reasoning approach to coordinated control of voltage and reactive power in order to enhance the voltage security of an electric power system is presented. The control strategy is expressed by simple rules, which measure the proximity of system state to certain operating conditions, and utilize linear equations to obtain the effective control models. The desired control actions are determined by considering several criteria at the same time. The procedure has been applied to a model system in order to verify its effectiveness. The simulation results show the advantages of this fuzzy modeling approach over conventional expert systems for voltage-reactive-power control. >

A critical comparison of search algorithms applied to the optimization of protein side-chain conformations

Abstract: Three different optimization algorithms are applied to solving the problem of finding the best side-chain conformations with a test set of 14 globular proteins having known crystallographic conformations. It is shown that simulated annealing, simple and modified genetic algorithms, and a heuristic combinatorial approach achieve similar optimal solutions, with the exception of simulated annealing applied to the largest proteins. The efficiency of the different algorithms, however, shows wide variations. General conclusions are drawn concerning the optimal approach to such problems. © 1993 John Wiley & Sons, Inc.

Including the load balancing issue in the optimization of multi-way join queries for shared-nothing database computers

Abstract: The authors determine a plan that makes the execution of an individual query very fast, with minimizing parallel execution time as the objective. This creates a circular dependence: a plan tree is needed for effective resource assignment, which is needed to estimate the parallel execution time, and which in turn is needed for the cost-based search for a good plan tree. A search heuristic that breaks the cycle by constructing the plan tree layer by layer in a bottom-up manner is proposed. To select nodes at the next level, the lower and upper bounds on the execution time for plans consistent with the decisions made so far are estimated and are used to guide the search. A query plan representation for intraoperator and interoperator parallelism, pipelining, and processor and memory assignment is proposed, as is a new approach for estimating the parallel execution time of a plan that considers sum and max of operators working sequentially and in parallel, respectively. >