https://ira.lib.polyu.edu.hk/bitstream/10397/1269/1/SetupSurvey-2006%28June-19%29.pdf

A survey of scheduling problems with setup times or costs

Benchmarking has become one of the most important methods for quantitative performance evaluation of processor and computer system designs. Benchmarking of modern multiprocessors such as chip multiprocessors is challenging because of their application domain, scalability and parallelism requirements. In my thesis, I have developed a methodology to design effective benchmark suites and demonstrated its effectiveness by developing and deploying a benchmark suite for evaluating multiprocessors. 
More specifically, this thesis includes several contributions. First, the thesis shows that a new benchmark suite for multiprocessors is needed because the behavior of modern parallel programs is significantly different from those represented by SPLASH-2, the most popular parallel benchmark suite developed over ten years ago. Second, the thesis quantitatively describes the requirements and characteristics of a set of multithreaded programs and their underlying technology trends. Third, the thesis presents a systematic approach to scale and select benchmark inputs with the goal of optimizing benchmarking accuracy subject to constrained execution or simulation time. Finally, the thesis describes a parallel benchmark suite called PARSEC for evaluating modern shared-memory multiprocessors. Since its initial release, PARSEC has been adopted by many architecture groups in both research and industry.

Benchmarking modern multiprocessors

https://web.stanford.edu/class/msande312/restricted/klos04facility.pdf

Facility location models for distribution system design

We describe an algorithm for the 0-1 knapsack problem (KP), which relies mainly on three new ideas. The first one is to focus on what we call the core of the problem, namely, a knapsack problem equivalent to KP, defined on a particular subset of the variables. The size of this core is usually a small fraction of the full problem size, and does not seem to increase with the latter. While the core cannot be identified without solving KP, a satisfactory approximation can be found by solving the associated linear program (LKP). The second new ingredient is a binary search-type procedure for solving LKP which, unlike earlier methods, does not require any ordering of the variables. The computational effort involved in this procedure is linear in the number of variables. Finally, the third new feature is a simple heuristic which under certain conditions finds an optimal solution with a probability that increases with the size of KP. Computational experience with an algorithm based on the above ideas, on several ...

An Algorithm for Large Zero-One Knapsack Problems

Publisher Summary This chapter presents a great variety of practical problems that are represented by a set of entities, each having an associated value, from which one or more subsets has to be selected in such a way that the sum of the values of the selected entities is maximized and some predefined conditions. The most common condition is obtained by associating a size to each entity and establishing that the sum of the entity sizes in each subset does not exceed some prefixed bound. These problems are generally called knapsack problems, as they recall the situation of a hitch-hiker having to fill up knapsack by selecting from among various possible objects those that provides the maximum comfort. The entities are called as items and their number is indicated by n. The value and size associated with the j-th item are called as profit and weight. The majority of problems are considered as single knapsack problems, where one container must be filled with an optimal subset of items. The chapter reviews both exact and approximate algorithms for the solution of the most important types of knapsack problems. Exact solutions are obtained through branch-and-bound algorithms or dynamic programming, or a combination of the two approaches. The performances of these algorithms are evaluated by computational experiments with randomly generated test problems. Approximate algorithms are of the greedy type or are based on scaling. They are evaluated on the basis of their worst-case performance.

Algorithms for Knapsack Problems

The classical generalized assignment problem (GAP) may be stated as finding a minimum-cost assignment of tasks to agents such that each task is assigned to exactly one agent and such that each agent's resource capacity is honored. This NP-hard problem has applications that include job scheduling, routing, loading for flexible manufacturing systems, and facility location. Due to the difficulty in solving "hard" GAPs to optimality, most recent papers either describe heuristic methods for generating "good" solutions or, in the case of optimizing methods, computational results are limited to 500 to 1,000 binary variables.In this paper we describe a special purpose branch-and-bound algorithm that utilizes linear programming cuts, feasible-solution generators, Lagrangean relaxation, and subgradient optimization. We present computational results for solving "hard" problems with up to 3,000 binary variables. An unanticipated benefit of the algorithm is its ability to generate good feasible solutions early in the process whose solution quality generally dominates the solutions generated by two recently published heuristics. Furthermore, the computation time required is often less than the time taken by the heuristics. Thus, we have an optimizing algorithm that can be used quite effectively as a heuristic when proof of optimality is not an absolute requirement.

Solving the Generalized Assignment Problem: An Optimizing and Heuristic Approach

In this note we present an efficient algorithm for the 0-1 knapsack problem and announce the availability of a callable FORTRAN subroutine which solves this problem. Computational results show that 50 variable problems can be solved in an average of 4 milliseconds and 200 variable problems in an average of 7 milliseconds on an IBM 360/91.

An Efficient Algorithm for the 0-1 Knapsack Problem

In this paper we consider the classical capacitated facility location problem. A branch and bound algorithm is presented which measurably improves upon the recent results of Akinc and Khumawala. The use of a specialized Lagrangean relaxation results in significantly tighter bounds than those for the traditional continuous relaxation. These bounds, when combined with penalties derived from the Lagrangean relaxation, enable many integer variables to be fixed at specific values. This results in fewer branches, and indeed for certain test problems taken from the literature, branching is not required. Average computation time for a battery of test problems from the literature has been reduced (conservatively) by a factor of 3.

Robert M. Nauss

Papers

Solving the Generalized Assignment Problem: An Optimizing and Heuristic Approach

An Efficient Algorithm for the 0-1 Knapsack Problem

An Improved Algorithm for the Capacitated Facility Location Problem

The 0–1 knapsack problem with multiple choice constraints☆

Optimal sequencing in the presence of setup times for tow/barge traffic through a river lock