Can Akkan

Network decomposition-based benchmark results for the discrete time-cost tradeoff problem

In project management, the project duration can often be compressed by accelerating some of its activities at an additional expense. This is the so-called time–cost tradeoff problem which has been extensively studied in the past. However, the discrete version of the problem which is of great practical relevance, did not receive much attention so far. Given a set of modes (time–cost pairs) for each activity, the objective of the discrete time–cost tradeoff problem is to select a mode for each activity so that the total cost is minimized while meeting a given project deadline. The discrete time–cost tradeoff problem is a strongly -hard optimization problem for general activity networks. In terms of what current state-of-art algorithms can do, instances with (depending on the structure of the network and the number of processing alternatives per activity) no more than 20–50 activities can be solved to optimality in reasonable amount of time. Hence, heuristics must be employed to solve larger instances. To evaluate such heuristics, lower bounds are needed. This paper provides lower and upper bounds using column generation techniques based on “network decomposition”. Furthermore, a computational study is provided to demonstrate that the presented bounds are tight and that large and hard instances can be solved in short run-time.

Task-failure-driven rebalancing of disassembly lines

Many reverse-logistics systems that collect and reprocess end-of-life products require a disassembly stage. The variability of incoming products and the extent of damage, which is more likely to occur during disassembly than assembly, create significant uncertainty in disassembly tasks, namely the possibility of failed tasks. Such failures may lead to some successor tasks being infeasible, which changes the work content of downstream stations. To improve the profitability of such a disassembly line, a mixed-integer-programming-based, predictive-reactive approach is proposed. In the first step, a predictive balance is created and then, in the second step, given a task failure, the tasks of the disassembled product with that task failure are re-selected and re-assigned to the stations (i.e. the line is rebalanced). In the second step, the objective function models both the profit obtained from the disassembled product and the possible increase in any stations workload beyond the predictive cycle time. Since this rebalancing approach affects the work content of stations, a discrete-event simulation study is also carried out to analyse and compare the performance of disassembly lines for optimally found line balances (predictive and reactive). The results show that, with the proposed approach, 24–32% of the monetary throughput lost due to not taking corrective action can be recovered.

Finite-capacity scheduling-based planning for revenue-based capacity management

Abstract Finite-capacity scheduling can be argued to be a crucial component of revenue-based capacity management. In that case, one way to plan production is to reserve portions of capacity for incoming customer orders as they arrive, in real-time. In such a planning method, the way these work-orders are scheduled affects the useable capacity, due to fragmentation of the time-line. Assuming the work-orders are rejected if they cannot be inserted into the existing schedule, we develop heuristics to minimise the present-value of the cost of rejecting orders and inventory holding cost due to early completion. We perform simulation experiments to compare the performance of these heuristics in addition to some common heuristics used in practice.

The two-machine flowshop total completion time problem: Improved lower bounds and a branch-and-bound algorithm

Abstract This paper presents a branch-and-bound algorithm for the two-machine flowshop scheduling problem with the objective of minimizing the sum of completion times. The main feature of the branch-and-bound algorithm is a new lower bounding scheme that is based on a network formulation of the problem. With extensive computational tests, we demonstrate that the branch-and-bound algorithm can solve problems with up to 60 (45) jobs, where processing times are uniformly distributed in the [1,10] ([1,100]) range.

Minimizing sum of completion times on a single machine with sequence-dependent family setup times

This paper presents a branch-and-bound (B&B) algorithm for minimizing the sum of completion times in a single-machine scheduling setting with sequence-dependent family setup times. The main feature of the B&B algorithm is a new lower bounding scheme that is based on a network formulation of the problem. With extensive computational tests, we demonstrate that the B&B algorithm can solve problems with up to 60 jobs and 12 families, where setup and processing times are uniformly distributed in various combinations of the [1,50] and [1,100] ranges.

Improving schedule stability in single-machine rescheduling for new operation insertion

The problem studied here entails inserting a new operation into an existing predictive schedule (preschedule) on a (non-preemptive) single machine by rescheduling its operations, so that the resultant schedule is the most stable one among schedules with minimal maximum tardiness. Stability is measured by the sum of absolute deviations of post-rescheduling start times from the pre-rescheduling start times. In addition to several simple heuristics, this study investigates a hybrid branch-and-bound/local-search algorithm. A large set of instances that include cases with inserted idle times allows for tests of the performance of the heuristics for preschedules with varying degrees of robustness. The results show that algorithms can be developed that significantly improve the stability of schedules with no degradation in Tmax. In addition, new insights emerge into the robustness characteristics of a preschedule. Specifically, the number of gaps in the schedule, equal distribution of total slack among these gaps, and the slack introduced beyond the amount enforced by release times all have effects on schedule robustness and stability.

A system for pricing the sales distribution from blockbusters to the long tail

The long tail of retailing has been both a challenge and an opportunity for online retailers. This article provides guidelines for enhanced decision making strategies in pricing dependent on popularity, cross-sales quantity and reservation prices. Our model shows that if customer willingness to pay, or reservation price, is higher for less popular items in a category, a unique optimal price path exists which requires deep discounts on popular items. However, if the reservation price is lower for less popular items, the optimal price path is conditional on the profitability of cross-selling and the potential loss from the business of loyal customers. Analyzing data on books, songs and movies from Amazon.com, we provide empirical support for our model findings. An analysis of the same set of movies available both as instant videos and DVDs allow us control for unobserved product characteristics and yields contradictory price paths along the sales rank distribution with increasing prices for DVDs and decreasing prices for streaming movies, as predicted by our model. Display Omitted We examine the optimal discounting strategy along the sales distribution by accounting for both own and cross-sales of items and the reservation prices in the implementation of a pricing DSS.Any non-decreasing reservation price function along the sales rank distribution yields an optimal discounting strategy in which retailers price hit items lower than any non-hit item.A decreasing reservation price along the sales rank distribution coupled with a significant potential loss from loyal customers and a lack of cross-selling potential leads to hit items priced higher than non-hit items.However, a decreasing reservation price, a negligible potential loss from loyal customers and an environment conducive to cross-selling may still give way to hit items priced lower than non-hit items.We analyze data on books, MP3 songs, DVD movies and instant movies from Amazon.com and show that the same set of movies available both as instant videos and DVDs have opposing price paths along the sales rank distribution with increasing prices for DVDs and decreasing prices for streaming movies, as predicted by our model.

Finding robust timetables for project presentations of student teams

This article describes a methodology developed to find robust solutions to a novel timetabling problem encountered during a course. The problem requires grouping student teams according to diversity/homogeneity criteria and assigning the groups to time-slots for presenting their project results. In this article, we develop a mixed integer programming (MIP) formulation of the problem and then solve it with CPLEX. Rather than simply using the optimal solution reported, we obtain a set of solutions provided by the solution pool feature of the solution engine. We then map these solutions to a network, in which each solution is a node and an edge represents the distance between a pair of solutions (as measured by the number of teams assigned to a different time slot in those solutions). Using a scenario-based exact robustness measure, we test a set of metrics to determine which ones can be used to heuristically rank the solutions in terms of their robustness measure. Using seven semesters’ worth of actual data, we analyze performances of the solution approach and the metrics. The results show that by using the solution pool feature, analysts can quickly obtain a set of Pareto-optimal solutions (with objective function value and the robustness measure as the two criteria). Furthermore, two of the heuristic metrics have strong rank correlation with the robustness measure (mostly above 0.80) making them quite suitable for use in the development of new heuristic search algorithms that can improve the solution pool.

A bi-criteria hybrid Genetic Algorithm with robustness objective for the course timetabling problem

Traditional methods of generating timetables may yield high-quality solutions, but they may not yield robust solutions that may easily be adapted to changing inputs. Incorporating late changes by making minimum modifications on the final timetable is an important need in many practical applications of timetabling. In this study, we focus on a subset of course timetabling problems, the curriculum-based timetabling problem. We first define a robustness measure for the problem, and then try to find a set of good solutions in terms of both penalty and robustness values. We model the problem as a bi-criteria optimization problem and solve it by a hybrid Multi-objective Genetic Algorithm, which makes use of Hill Climbing and Simulated Annealing algorithms in addition to the standard Genetic Algorithm approach. The algorithm is tested on the well known ITC-2007 instances and shown to identify high quality Pareto fronts.

How Trending Status and Online Ratings Affect Prices of Homogeneous Products

ABSTRACT The online retailing environment has grown more complex with the myriad information items available for customers to peruse. Two such information items include the trending lists of currently popular products, and online customer reviews and ratings of products. A trending product is one that relatively large groups of people are currently browsing, purchasing, or discussing. In this paper, we analyze the pricing implications of both types of information items. Our main goal is to test whether, acting as signals of desirability and popularity, a trending status or high ratings and reviews can lead to price cuts on homogeneous products. We provide empirical support for a model that demonstrates these relationships using data on books from twenty-four categories of Amazon.com. An important theoretical contribution of our model is that even in the absence of strategic considerations or demand dependencies, our model suggests lower prices for trending and positively reviewed products. Our work shows that pricing systems using popularity and average customer ratings as inputs can help achieve increased profitability.