This study focuses on a new real-word problem encountered in the construction sector, which concerns the optimization of the removal of construction waste bins from construction sites to a massification platform, where a limited heterogeneous fleet of tipper trucks (vehicles) must perform direct trips from the platform to the construction sites to collect the waste bins. Each vehicle has a capacity of one bin, it leaves the platform with an empty bin, travels to the construction site, drops off the empty bin in the construction site, collects the full bin and returns to the platform to unload the full bin. The issue is that the vehicles and the construction sites have one or more periods of availability, and thus are not available any time. This problem is modeled as a parallel machine scheduling problem of bin removal tasks on non-identical machines (vehicles), with new constraints that concern the presence of multiple availability intervals for both vehicles and tasks. Two mixed-integer programming (MIP) models are presented and evaluated on 18 new instances derived from real industrial case study.
Safae Abderebbi, Wahiba Ramdane Cherif-Khettaf. Optimization of Direct Transportation Flows for the Removal of Construction Waste Bins with both Resource and Task Availability Interval Constraints. The optimization of the Big-bag removal is studied in (Jaballah and Ramdane Cherif-Khettaf, 2021; Ramdane Cherif-Khettaf et al., 2022) where a new vehicle routing model called the Multi-Trip Pickup and Delivery Problem, with Split loads, Profits and Multiple Time Windows was proposed. This model allowed mutualization of material delivery with Big- Bag waste removal, using tail-lift truck fleet. In this study, we focus only on waste bin removal, which consists in performing a set of direct trips from the platform to the construction sites to satisfy construction site requests for full bins removal and their replacement by empty bins. The remainder of the paper is organized as follows. Section 2 describes the problem and related literature; two MIP models are presented in section 3. Experimental results with the definition of benchmarks are given in section 4. Finally, concluding remarks are given in section 5. 2 PROBLEM DESCRIPTION AND RELATED LITERATURE A limited heterogeneous fleet of tipper trucks (vehicles) situated at the platform must perform multiple direct trips between the platform and the construction sites. The vehicles differ by their speed and their distance limit. Each trip consists of loading an empty bin at the platform, delivering it to a given construction site, collecting a full waste bin from this construction site, and unloading it in the recycling center located just next to the platform. Each vehicle has one or more periods of availability, which represent the time windows when the vehicles are available at the platform and so can move to the construction sites to collect waste bins, outside of these periods of availability, the vehicles can be mobilized for other tasks external to the platform and can’t satisfy the request of bin removal. The platform also has a period of availability, that is given by the platform's opening hours. The construction sites have one or more types of bins depending on the specificity of the works in progress in the construction site (bin for wood, bin for metals, bin for plaster, bin for inert materials, etc.). Each site has only one bin per type, and must therefore send to the platform requests for bin removal. The platform must manage the bins waste removal with the available resources and ensure the replacement of each full bin with an empty one. The construction sites also have one or more periods of availability during which vehicles access is allowed, and thus the arrival and departure of vehicles at the construction site location must be within one of the availability time windows allowed by the construction site. We denote by a bin removal task, all operations that consists of loading the vehicle with an empty bin at the platform, travelling from the platform to the construction site, unloading the empty bin at the construction site, loading the full bin, travelling back to the platform and unloading the full bin at the platform's recycling center. Platform service time is the time required for loading and unloading the bin on the platform. This problem can be modelled as unrelated parallel machine scheduling problem in which the vehicles can be represented as machines with multiple periods of availability and the bin removal tasks as jobs that have one or multiple periods of availability, and require a certain amount of processing time, that depends on the vehicle that is used. In addition, the constraint of availability of the tasks in our case concerns only a part of the task processing time, it is the loading and unloading part at the location of the construction site and does not concern the part of the travel to and from the site. The objective is to perform the maximum number of bin removal tasks, to determine the assignment of tasks to the availability intervals of the vehicles; and to define the sequence of satisfied tasks per available interval of each used vehicle. In terms of computational complexity (Lenstra et al., 1977) proved that the single machine scheduling problem with only release dates, which is a special case of our problem is NP-hard. In literature, extensive studies have been conducted in the area of parallel machine scheduling with time constraints without availability constraints (Arik et al.,2022; Osorio-Valenzuela et al., 2019). In most of the research reviewed in the area of parallel machine scheduling, the availability constraints are defined on resources (Such-Jeng, 2013). Very few studies consider the availability intervals of tasks as in (Gedik et al.,2016). A survey on parallel machine scheduling under availability constraints can be found (Kaabi and Harrath, 2014). Despite the abundant literature on parallel machine scheduling, the problem that we present here is in our knowledge a novel one and allows us to model a new constraint in unrelated parallel machine scheduling problems, that is both resource and task multiple availability interval constraint. Our contribution can be summarized in the two following issues: ─ Modeling a real problem of direct transportation of bin waste in the construction sector as a parallel machine scheduling problem with a new specific constraint that is multiple periods ofICORES 2023 - 12th International Conference on Operations Research and Enterprise Systems, Feb 2023, Lisbon, Portugal. pp.221-228, ⟨10.5220/0011821900003396⟩. ⟨hal-04056131⟩
