AI in Logistics: The Ultimate Disrupter; How to Manage Last Mile Costs AI in Logistics: The Ultimate Disrupter; How to Manage Last Mile Costs
Over half the total cost incurred by Logistics companies falls in the execution of first and last-mile services.  Logistics companies are... AI in Logistics: The Ultimate Disrupter; How to Manage Last Mile Costs

Over half the total cost incurred by Logistics companies falls in the execution of first and last-mile services.  Logistics companies are seeking to reduce costs by making the transportation process faster, more reliable, and more efficient.  With the growing digitization of the logistics industry, more and more companies are adding artificial intelligence (AI) to their logistics and  supply chain in order to maximize their resources by reducing the time and money spent on figuring out how, where, and when to deliver  goods and services.

Artificial intelligence (AI) is the simulation of human intelligence processes by machines, especially computer systems. Artificial Intelligence can be used to improve logistics experience by increasing reliability, reducing the cost of transportation, faster processing, and deciding optimal routes for last-mile operations.  Given the intense competition in the industry, customer satisfaction has become a key battleground for business dominance; companies like Amazon and Alibaba are investing hundreds of millions of dollars  in order to make their delivery processes that much faster and more efficient to ensure the next day or even same-day delivery  Artificial Intelligence (AI) is the primary tool used to achieve these objectives by finding the best route, estimating accurate delivery times as well as determining the correct warehouse product mix while minimizing costs.

[Related article: Redefining Robotics: Next Generation Warehouses]

AI in Logistics

Nowhere are efficiencies introduced by AI more visible than in the areas of vehicle routing, network planning, and predictive demand.  Machine learning led innovations in these areas have made businesses more agile and dynamic permitting Uber and Grab to deliver outstanding customer experience.  Better capacity planning, optimal route planning, dynamic charging, optimal allocation of resources and vehicles quickly to the in-demand areas to reduce customer wait times are just some examples of how AI is used today.

ai in logisticsFigure 1: AI has been called the secret sauce of logistics. But just how does AI help? We have highlighted seven applications of machine learning which are widely used in Logistics.

In this article, we will see how vehicle routing problem (VRP) algorithms can help Logistics companies make significant operations savings.  Vehicle routing is a complex logistics management problem and represents a key class of problems to solve in order to lower costs and optimize logistics resources.  Route planning has several uncertainties – volatile demand for service, traffic incidents, unexpected weather conditions etc; all of these uncertainties can lead to greater costs in logistics management if not taken into consideration properly. The main purpose of VRP is to find optimal routes for multiple vehicles visiting a set of locations by minimizing both time and cost.

Rosebay works with several regional logistics industry partners and customers, including Post Office Indonesia (POS Indonesia) in making their organization data-driven and Vehicle routing optimization is one of the most common use cases that we encounter in Logistics.  And for a good reason – it has the potential to make the maximum impact on a Logistics company’s bottom line by solving some of the most expensive first mile and last mile service problems and in the process represents a significant optimization that logistics businesses need in order to compete effectively.

ai in logisticsFigure 2:  Postmen at POS Indonesia heading out from the depot for delivery.  First mile and Last mile services account for over half the cost of Logistics companies.

Last-mile/First Mile services optimization via Vehicle Routing Optimization

The Vehicle Routing optimization (VRO) has various applications in real life. The main objective of this problem is to design routes for vehicles that depart from a given number of different depots, need to go through several locations to deliver some service, and once the shift is over they return to a set location. The complexity of the operations differs by customer; you can add your business conditions, such as the load capacity of vehicles, the maximum distance they can travel per day or the duration of the working shift of your drivers.  The goal of VRO given these conditions is to compute a route which minimizes the aggregate transport costs such as the total distance traveled number of vehicles used and/or the total transport time.  

Figure 3: VRO is important for last mile and first mile optimizations.

VRO can also be classified as single-vehicle and multi-vehicle.  The single VRO case consists of optimization for a single vehicle with a single start location depot and a single end location – the optimization consists of finding the correct sequence of addresses to deliver services in which minimizes cost. The Multi-vehicle VRO is a generalized version of the above whereby multiple vehicles are involved with multiple start and endpoints and with several intermediate locations in between.  The solution, in this case, consists of finding the optimal routes for all vehicles given business constraints while minimizing cost.

Making a comprehensive optimization roadmap for Logistics companies

Recently we partnered with POS Indonesia to provide a comprehensive vehicle routing optimization solution, starting from individual single-vehicle route optimization to countrywide multi-channel end-to-end logistics optimization.  Based on the work we have performed with several regional companies in this sphere in the past few months our results indicate total savings of anywhere from 22% to 71% in terms of time, distance and fuel expenses when implementing the complete three-phase roadmap as illustrated in Figure 4.  

Figure 4: A road map for logistics companies with a national or international footprint. Benefits can come in the form of efficiencies of cost as well as increased customer satisfaction. Altogether the three-phase taken together bring the highest set of benefits to Logistics companies.

For a logistics company with a nationwide or international footprint, we typically subscribe to a three-phased incremental approach – with resources and allocations for each phase based on the results of the previous phase:

  • Phase 1.  Consists of simple local improvements in single-vehicle routing.  This is the simplest phase but captures some of the highest value (anywhere from 15% for companies that are already using efficient vehicle routing protocols to 60% in companies which do not have optimized operations).  The results of Phase 1 also serves as a benchmark for improvements in the company for Phase 2 and Phase 3  
  • Phase 2 consists of extending the single-vehicle to multi-vehicle routing and fleet management for a certain locale.  Savings in time and money of over 20% are not atypical in phase 2. Taken together with Phase 1 we have seen optimizations of up to 47%.  
  • Phase 3 consists of combining the vehicle routing with capacity planning and demand prediction to make a comprehensive end-to-end routing platform for goods and services.  While savings may not look at significant compared to Phases 1 and 2, Phase 3 benefits accrue more in the form of customer satisfaction. Faster end-to-end delivery, tracking and prediction is one of the surest ways to increase satisfaction of customers and gain an edge over competitors. 


Case study:  Savings from single-vehicle route optimization in Bandung and Jakarta city

Rosebay recently conducted field trials of its vehicle routing optimization solution in Bandung and Jakarta.  Twenty sets of eight locations were chosen both in Jakarta and in Bandung. Figure X below is a single example from the twenty sets from Bandung.  The noted markers represent real delivery locations of our partner logistics company. It should be noted that the un-optimized/control case represents the unsolicited route taken by delivery personnel.   

Figure 5: Results of a single VRO solution for the city of Bandung. Twenty such sets were field-tested in Jakarta and Bandung with great results. Phase 2 and Phase 3 are now underway.

Average data from twenty locations indicate savings of around 48% in time and 31% in distance as a direct result of the optimization.  Our client is expected to save over $1.7 Million in fuel costs alone as a result of introducing these efficiencies in its nationwide fleet.

[Related article: Trends in AI: Towards Learning Systems That Require Less Annotation]

Given the positive results of Phase 1, Phase 2 and Phase 3 tests are underway and are expected to give similarly encouraging results.

Planning your AI/Digital Transformation journey

Many organizations have now been benefited with investments in artificial intelligence.  While in western nations 15% have already started to use AI while other 31% plans to have them implemented in 2019, the figures are much lower in Asia (less than 5%).  Rosebay is planning to hold workshops in cities around Asia in December and January, including in Jakarta, Kuala Lumpur, Ho Chi Minh City, and Phnom Penh to help companies understand the importance of AI and Big Data.  There are three key ways where AI can contribute to commercial and business success:

  1. AI for Business growth
  2. AI for Optimizations/savings 
  3. AI for Managing risk   

Contact Rosebay at sujata.sharma@rosebayconsult.com to get a special invite to our AI workshop if you are in Jakarta, Kuala Lumpur, Ho Chi Minh City or Phnom Penh.

Sujata Sharma

Sujata Sharma is a Business Analyst at Rosebay Consulting. She is working on Artificial Intelligence and Big Data projects. Sujata has 4+ years of experience working as a data analyst for the financial services and audit industry. Sujata holds a Bachelor’s degree in Business Administration. Currently, she is working towards a Master’s degree in Economics.