Yunsong Guo

Multiple crossdocks with inventory and time windows

Crossdocking studies have mostly been concerned with the physical layout of a crossdock or on a single crossdock. In this work, we study a network of crossdocks taking into consideration delivery and pickup time windows, warehouse capacities and inventory-handling costs. Because of the complexity of the problem, local search techniques are developed and used with simulated annealing and tabu search heuristics. Extensive experiments were conducted and results show that the heuristics outperform CPLEX, providing solutions in realistic timescales.

Heuristics for a bidding problem

In this paper, we study a bidding problem which can be modeled as a set packing problem. A simulated annealing heuristic with three local moves, including an embedded branch-and-bound move, is developed for the problem. We compared the heuristic with the CPLEX 8.0 solver and the current best non-exact method, Casanova, using the standard CATS benchmark and other realistic test sets. Results show that the heuristic outperforms CPLEX and Casanova.

Carrier assignment models in transportation procurement

This paper extends carrier assignment models used in winner determination auctions for transportation procurement to include shipper non-price objectives and carrier transit point costs. The models are unlike traditional carrier assignment models which incorporate only carrier lane bids, and different from combinatorial auction models which focus on packets of lanes without considering transit point costs. We develop solutions, including metaheuristics, for the new models and through computational experimentation show that the algorithms work well and can be easily implemented.

Using a Lagrangian heuristic for a combinatorial auction problem

In this paper, a combinatorial auction problem is modeled as a NP-complete set packing problem and a Lagrangian relaxation based heuristic algorithm is proposed. Extensive experiments are conducted using benchmark CATS test sets and more complex test sets. The algorithm provides optimal solutions for most test sets and is always 1%from the optimal solutions for all CATS test sets. Comparisons with CPLEX 8.0 are also provided, which show that the algorithm provides good solutions

A Non-Exact Approach and Experiment Studies on the Combinatorial Auction Problem

In this paper we formulate a combinatorial auction brokering problem as a set packing problem and apply a simulated annealing heuristic with hybrid local moves to solve the problem. We study the existing exact and non-exact approaches to the problem and analyze the performance of those approaches. We compared our heuristic with the leading exact method CPLEX 8.0 solver and another non-exact algorithms Casanova using both the CATS test sets and test cases believed more difficult than CATS. Results show that the method is competitive with CPLEX 8.0 and obtains near optimal solutions for the CATS cases and up to 15% and 40% better solutions compared with CPLEX and Casanova, respectively, when the other instances were used.

Machine scheduling performance with maintenance and failure

In manufacturing control, machine scheduling research has mostly dealt with problems either without maintenance or with deterministic maintenance when no failure can occur. This can be unrealistic in practical settings. In this work, an experimental model is developed to evaluate the effect of corrective and preventive maintenance schemes on scheduling performance in the presence of machine failure where the scheduling objective is to minimize schedule duration. We show that neither scheme is clearly superior, but that the applicability of each depends on several system parameters as well as the scheduling environment itself. Further, we show that parameter values can be chosen for which preventive maintenance does better than corrective maintenance. The results provided in this study can be useful to practitioners and to system or machine administrators in manufacturing and elsewhere.

MINIMIZING THE MAKESPAN FOR UNRELATED PARALLEL MACHINES

In this paper, we study the unrelated parallel machine problem for minimizing the makespan, which is NP-hard. We used Simulated Annealing (SA) and Tabu Search (TS) with Neighborhood Search (NS) based on the structure of the problem. We also used a modified SA algorithm, which gives better results than the traditional SA and developed an effective heuristic for the problem: Squeaky Wheel Optimization (SWO) hybrid with TS. Experimental results average 2.52% from the lower bound and are within acceptable timescales improving current best results for the problem.

Minimizing total flow time in single machine environment with release time: an experimental analysis

We study the problem of minimizing total flow time on a single machine with job release times. This problem is NP-complete for which is no constant ratio approximation algorithm. Our objective is to study experimentally how well, on average, the problem can be solved. The algorithm we use produces non-preemptive schedules converted from preemptive ones. We evaluate average solution quality for the problem to identify the characteristics of difficult instances. Results obtained are compared with those recently obtained by other researchers. Based on extensive experiments, we also develop an empirical model to predict solution quality using interpolation.

Tariff concessions in production sourcing

In this paper, we study a multi-stage production sourcing problem where tariff concessions can be exploited at the firm level using free trade agreements between countries. To solve the problem, an algorithm which embeds a very large-scale neighborhood (VSLN) search into a simulated annealing framework is developed. A numerical study is conducted to verify the effectiveness of the solution approach.

ExOpaque: A Framework to Explain Opaque Machine Learning Models Using Inductive Logic Programming

In this paper we developed an Inductive Logic Programming (ILP) based framework ExOpaque that is able to extract a set of Horn clauses from an arbitrary opaque machine learning model, to describe the behavior of the opaque model with high fidelity while maintaining the simplicity of the Horn clauses for human interpretations.